Video details

Introduction to Python for Automating Business & Financial Processes with Liberty Mutual

Python
10.31.2021
English

Presented by Women Who Code Python and Liberty Mutual Insurance 👩‍💻 Speaker: Kate Ferguson, Moderated by Stephanie Rideout ✨Topic: Introduction to Python for Automating Business & Financial Processes with Liberty Mutual
In this talk we will learn how to identify opportunities to automate manual, repetitive processes for finance/business peers as well as understand how to actually automate them using basic Python. Kate will share her experiences working with finance and business partners and how she has identified opportunities to automate monthly or daily processes that they perform manually. Since technology isn’t their full-time job, she has both introduced them to a new way of approaching their work and empowered them to focus on the analysis for the customer, rather than the process.
Self-Learning & Helpful Tips: https://www.codecademy.com/learn/learn-python-3 https://datatofish.com/python-tutorials/
https://www.learnpython.org/en/Welcome https://www.w3schools.com/python/ Getting Started:
Download Python: https://www.python.org/
Python Built-in Features: https://docs.python.org/3/library/
For our 💬 slack channel, 🎥 previous event recordings, 🗓 upcoming events, 💻 github repo and more check us out on https://beacons.ai/wwcodepython
___ 💻 WWCode Digital Events: https://www.womenwhocode.com/events​ 🔎 Job Board: https://www.womenwhocode.com/jobs​​​​ 💌 Make a Donation: https://www.womenwhocode.com/donate

Transcript

So welcome, everyone, to our introduction to Python for automating business and financial processes with Kate Ferguson, who is associate data analyst at Liberty Mutual. And I want to give a big thank you to Liberty Mutual today for sponsoring this fantastic event. We're so happy that you could sponsor it today. Thank you so much and a little bit about our speaker today. Kate is an associate data analyst analyst at Liberty Mutual. She has one and a half years worth of experience working on a cloud finishing Ops team and helps manage public cloud spend and optimization for the company. She loves to be a bridge between finance and technology and has degrees in both finance and mis. When she is not traveling around the world, you can catch her editing academic papers and eating chocolate ice cream. That's wonderful. So without further Ado, I will now pass the session on to Kate. Awesome. Thank you, Stephanie. Hi, everyone. Welcome. Bear with me while I share the presentation. Always the hardest part. All right, let's see. All right. Awesome. So, Hi, everyone. Thank you, Stephanie, again and welcome. I'm super excited to be here today to talk about how we can leverage simple Python process. Simple Python right to improve business and financial processes in your organization and talk about how I've done it here at Liberty Mutual. I'm sure if you've ever worked with Excel, you may have noticed it's not very good at handling the excessive data and tends to crash quite often. But by using coding languages like Python, you can work to avoid crashes and limit human error. The key to this happening, though, lies in the creation of a bridge between business and technology. Personally, I believe that this is the future of work in the industry, and it will lead to more automation of routine tasks and increased analysis being done by humans. I hope that by the end of this webinar, you'll have a good sense of how you can apply Python to get ahead in your career and provide more value to your organization. So let's get started. Here's a quick overview of what I attempt to cover today in this webinar. We're going to start with a brief introduction to Liberty Mutual and myself. We'll go over why we should build a bridge between business and technology. Then we'll look at using Python and business and some good places to start. If it's your first time, then I'll go over examples of how I've used Python here at Liberty Mutual to create efficiencies and show two demos of those in action. Finally, we'll discuss how you can take what we've talked about here today and apply it to your career. Let's get to it. Liberty Mutual is currently the 6th largest property and casualty insurer in the world. Technology underpins everything we do and data plays a huge role in ensuring that all of our customers are taken care of. And that's the reason why we have over 5000 technology employees. We've also been named one of ITG's Insider Pro and Computer Roles Best Places to Work in It for the second consecutive year, we climbed 16 spots since 2020, and we're really, really proud of that. Personally, I'm thankful to be at a place that has been listed here because it's really been helpful in developing my technical skills and knowledge that I hope to share with you today. Over the past year, we've also been recognized by multiple organizations for our continued commitment to our employees. Notably, we've been recognized as one of Forbes Best Employers for Women and for New Grads. Now, both of those I can personally attest to the company being great at it's been really good so far. And then finally, for Liberty here we have a group called Women in Technology that I'm proud to be a part of. I'm currently a member of our recruiting and retention work stream, and we focus on improving the employee and a candidate experience to help attract more women to technology roles and keep them there. We also have other work streams, including Legit, which works to close the gender gap by working with girls to spark their interest in Stem and technology from an early age. I really enjoyed working with Wit so far. And if you want to learn more about what we do, feel free to reach out. If you have any other questions about Liberty or technology at Liberty Mutual, feel free to go to this link to learn more or reach out to me personally. So now a little bit about myself. So I'm based out of Boston, Massachusetts, and I'm currently an associate data analyst for Clouds and Ops Team. My first role out of College, though, is as an associate financial analyst where I worked in the tech finance Department of our company Fun Facts About Me I'm a frequent traveler. The picture on the right is actually taken in Peru in case Alamos on my sweater didn't give it away. And I'm also an avid crossword and puzzle enthusiast. You can catch me at Trivia nights most Wednesdays, and I also try the crossword every day from New York Times. I only do a couple of days a week. Well, but I try, and if you ever want to reach out here some links, they'll be at the end, too. So let's talk a little bit more about my journey and what led me to be talking with you today. I went to College at upstate New York and majored in finance with a minor in mis or management information system, it was there that I began learning how coding could be used in a business context, and even though I am by no means a master coder and we're so dabble in it, I've been really proud of and happy with the impact that it's had so far in my career in such a short amount of time. During my time at RIT, I did a summer internship with Mozilla in San Francisco, where I was a financial analyst intern. This led me to have the opportunity to work with it real estate and technology innovation apartments to help them manage their spend. It was there that I learned I really wanted to work as an intermediary between finance and technology later in my career. Once I graduated, once I learned this, Liberty Mutual somehow had the perfect job posting. When I was looking financial analysts in technology finance, it was perfect. It was everything I was looking for. And when I started working here, I was placed in the technology Finance Department, which partnered with teams in the enterprise to manage their hosting and server technical spend. However, I was specifically assigned to the Cloud Thin Ops team. This is a group that works across the enterprise to manage spend in the public cloud millions of dollars to spend across AWS, Azure and GCP. Those are the big players. We also create optimization awareness for the entire enterprise, and we help provide cloud education. This means that we're really at the middle between business and technology, right? Because we have to do the financial processes each month to account for the cloud spend. But we also almost act as consultants for the business stakeholders. We partner with them. We help them break down finance terms and technical terms, and we break down the technical terms from the cloud to the finance peers that we have. Right? So we're really right at the intersection. And so you'll notice here that Cloud Ben Apps has two departments. But I worked with Claudin Apps the whole time. That's because earlier this year closing in Ops team moved to the Secure DevOps or SDP Department. This is because they enable developers to easily deploy their apps through pipelines, report on the status of apps in the cloud, and so much more so by moving from technology finance to this organization, we're able to put ourselves at the heart of the developer experience and sort of embed our cost, transparency and effort right where the action is. With this niche positioning, I've been fortunate to have partnered with full time software engineers, architects, business analysts, financial analysts, and so much more in order to get my job done. And as a result of these two roles and partnerships, I've really seen how beneficial new technology can be for improving business processes. In addition to the learning curve engineers have as they learn the business processes. So now I want to sort of share the challenges and the outcomes as a result of building that bridge here at Liberty Mutual. So why build a bridge? Right? Why just swim like you need a bridge? Well, currently, in my opinion, the business and technology Department are oftentimes too siloed in corporate organization, so let's walk through it. Business teams often have standard, routine processes and rely on Excel pretty heavily. And that's not to say Excel isn't great. I love Excel. I come from Finance I live in Excel most of the time, but it definitely comes with its flaws. Right? It can't have a large data and has limited opportunities for automation. Right. Like really good automation. Even though business people may be aware of non Excel solutions, they don't learn how to code in College. And normally Excel is good enough. Right. You can get a good enough to get the job done and be on your Merry way. But they can also feel lost if they say, okay, maybe Excel keeps crashing 20 times in 20 minutes. What's a better way if they want a different way, they often feel lost in knowing where to start looking because they don't learn how to code in College. Right. The technology teams are really great at creating quick, logic based solutions that use code. But the problem is they're a non exposed to business teams and the processes they do very often and B normally they don't learn how to do accrual accounting in College. Right. It's just not part of the curriculum. So there's an almost cognitive dissonance between these departments where there could be really beneficial synergies. So by partnering and working together, efficiencies come created and there's more time for value added work for the organization on both ends. As a result of this, information is more easily accessible and quicker, and as a result, it's easier to make decisions. You can make better informed decisions with more information. So how can we start to build this bridge and bring together both business and technology? A great place to begin is by partnering and creating solutions. Python okay. You may be saying that's great, but why Python? There are so many languages out there, right? Python is great because it's an open source language. This means it's free and there are tons of resources online. So if you're a business person like I am, who wants to become a little bit more technical, Python is great because it has a really easy learning curve and can be a quote, unquote gateway coding language. Right. So once you learn certificate logic, how it works, it can lead to maybe R or C Plus Plus or whatever language you want to jump to next. Right. It can be used to automate a lot of Excel process fees or other standard work. And I think the impact Python can have on Excel is similar to how Excel had an impact on calculators. Right. Well, you can do a lot of things with Calculator, especially the fancy ones. Nowadays. You can do a lot of things more quickly and efficiently by using Excel. To me, the analogy is Python from Excel, you can do so much more, and it's so powerful. Excel is so good, it still works. But Python is really exponentially better, in my opinion. And since it's open source, that means that people out in the world are actively working on creating their own packages with our sets of tools, features, capabilities to fit their needs. Currently, when I checked earlier this week, there's over 330,000 packages out there. So if you need something to do, I'm sure there's at least one, if not 100, that will suit your needs. And that can be daunting, right. 330,000 packages? Where do I even start? If you ever have questions on writing code or where to start? I found Stack overflow is a great resource. It's an online site that has millions of coding questions and answers. In fact, if you look at the picture on the right, you can see, as of July, there was over 1.8 million questions asked tagged with Python, right. And that's probably just the only ones that were tagged. There's probably more that haven't been. And when I checked earlier this week or earlier today, rather, there's 6700 questions tagged with Python asked in a week. That's a lot. So if you have a question about Python, the language, the phrasing, the punctuation, someone out there can help you get it done if it hasn't been asked already. And if you look at the bottom right. I know it's really small, but three of the related Tags there are actually packages that I hope to go over with you in a bit more depth. These packages are Pandas, NumPy and Tikinter. These are some of the best packages I've come across in getting started with using Python for business. So let's dive in Pandas. No, it is not a Panda on a computer. Unfortunately, that would be really fun. What it does do is it's a package that is great for beginning to explore the opportunities Python has because you can do so much and it's relatively easy to learn. It's also one of the most popular ones, right? It's not stack overflow. I think it has the most related Tags questions to it. With panels. You can bring in multiple Excel sheet, compare, manipulate, analyze data. You can format data and do some simple machine learning. If you really want to get fancy with it. You can also use this package, for example, to analyze stock trends or something like that, or even emissions data from different countries. It's really popular with finance, statistics and engineering careers. I can attest that. Finance is really popular because it can do so much with Excel. By no means have I done everything Pandas has to offer, but I found it really helpful when I want to automate something from Excel. It's pretty good at doing what I needed to do. Next. We have NumPy. Numpy stands for numerical Python. So this is a great package. If you're looking to do anything computationally, it's really popular in the data science community, and it's really good at performing analysis. It creates and analyzes a rate. Think of it as like a list, but it's much quicker than using the standard Python list feature. And you may say, well, why it was so different. It's because it stores the arrays locally and all in one place. Right. So if you have multiple lists, it can quickly look up and calculate and give you that result right away. You can do this for basic algebra statistics. Right. So min Max mode all that fun stuff, and you can also use it to ensure data quality. Look for missing values or base missing values, et cetera, et cetera. So this one is really good for anything numbers. That's what I like to think of it for. And then Tinter is a package that allows you to create a graphical user interface or GUI and Python. It's built into Python standard library packages. And it's really great if you want to make an interactive report, especially that involves calculations of some kind. Right. So I've seen examples. I remember College. I did an example with Tinter building a sales report. So if you're a salesperson filling out sales order form and you're a manager and you want to see all the sales, it's really popular because it doesn't require a command line interface. In order to interact with it, you can go press buttons, menus, pop up boxes, all that stuff. So it's really easy to streamline and make it easy for the user to understand. So enough about the packages I recommend using. Let's now look at how I use these packages to create Efficiencies during my time of Liberty Mutual. So the first example is one that was given to me in the first month of my career. Here, a finance product owner requested that I am tied for 800 cloud reservation purchases, but let's take a step back. What does that even mean? Amortize 800. What does that mean? So to amortize in finance, it's an accounting principle where you write down the value of an asset over time. So, for example, say you buy a $50,000 car and you expect it to last five years each year you would advertise or write down the value of the car by $10,000. Relatively simple. Except in this example, we're talking about cloud reservations or an upfront payment for one year worth of cloud usage instead of a car, and there's 800 of them. So the product owner wanted to see the cost by day over the lifetime of the reservation. And after getting the assignment, why didn't I just do it in Excel? I tried. Why did I search for an automated solution? Because Excel broke. It was so much data. We had 800 rows and think of it. We have to add 365 rows after each transaction, divide the purchase costs by 365. Copy and paste that 364 times, and then repeat that 800 times more than 800. Actually, it's a lot of work, and I still didn't like how much data and computing was required, especially for such a simple assignment. But it should be a simple assignment, right? Additionally, it would have been a very time consuming TDs and error prone. What if I got tired and I inserted 362 rows instead of 365. How would I know? How would I fix that? There's 800 of them. Right. Service for little errors like that was really high. So I really liked the idea of some coding the work instead of me. So what did I do? So in Python, I wrote a script taking the file. The Excel file had 365 rows in between each transaction line. Copy the transaction details for the 364 rows and perform the calculation. For me to do this, I use NumPy from the numbers. Right. And I use Pandas to manipulate the Excel. The outcomes of this for the project that could have taken days or three weeks. I was a bit concerned how long it was going to take. It was done in less than an hour. Right. So huge savings efficiencies. Think of it from the business side. Huge costing, right. Much less compute energy needed less time or more time that I could be doing other things. But that's sort of where I stopped looking at the outcome. I more or less just gave the product owner exactly what they had asked for and didn't do anything else. Now, if you think about it, 800 365. That's like over 300,000 rows. Probably not the best customer service I could have given. Right. So this going further section is really thinking. Okay, how could I have made this better and push the code a little bit further? I could have made a monthly report each month, and I give them exactly what they need to perform their book close analysis. So each month they have to record how much of the reservation was used, which is why they needed it by day. And instead of making them query the data or deal with that massive Excel file, I could create a report each month for them using Python. So that's what I did. Let's now see how I did it. And what I'm going to do is I'm going to share the code and the files. And I just want you to know that all of the data that I'm sharing today is dummy anonymized data, so can't steal our reservation. All right. First, this is the original file. I was given the CSV. And what really matters here is the period. So this is the date it was purchased or will be purchased, I guess. And this is the cost. So this is how much we're looking to advertise. For example, it was purchased on October 29, and we need to amortize $365 over 365 days. So one dollars a day. Right. Let's see if we can do it. We open up the code what I first did and let me know, Stephanie, if this is too small, but what I first did is important. The Pandas and NumPy packages from Python. Yeah. Could you please Zoom in a bit? Is that better? Can you Zoom in just a table? Better. Perfect. Thank you so much. Okay. Thank you. So what we did first is important, the Pandas and NumPy packages that we needed. And then we brought in the file that we need. So this Pandas TD read CSV is bringing in the file. So if we run it and then all this is going to do, we're going to double check that the data looks like it's supposed to remember the Excel file. So we're going to run it and we can see perfect. You see, 1029, 365. Everything else is looking good. The only thing that's really different here is this numbering over here. Right. So instead of the Excel rows, what it is, it created a data frame, and these are just the labels or the numbers or indexes, rather, for each transaction detail or transaction row. So next, what we're going to do is we look at the cost 365 point right now. It's sort of a text. Right. What we need in our two calculations is we need it as a number. So what we're going to do is we're going to convert the blended cost or the upfront cost, and we're going to convert it to a float, which basically is Python for number. So we're going to run that and you can see now the only thing that really changed is the decimal. How many decimals there are. Everything else is the same. So we're golden. Next, we're going to reformat the text of the date. We're going to change it to a date so that we can manipulate it. We need to recognize that over 365 days. So instead of typing in 10, 29, 30, we're going to have this do it for us if we run this or let's see the before. Sorry, 1029 all the slashes. All it's going to do is change it to Dashes and change the data type. So here's Dashes reorganized it. And now Python is able to say that's the date next. What we're going to do is we're going to divide the purchase cost by $365. So the 365 better go to one. Otherwise something is broken. So if we run that, we'll see it worked. Perfect. We can see the one here. The next transaction is also lower and so on and so forth. But the problem is there's only one of them. We need 365 in order to account for this. Right. What we're going to do is using NumPy. What this code is saying is basically add 365 rows and use a forward bill. So that means we're going to add 365 rows in between each, like zero in between the one, two, so on and so forth for all 864 rows of reservations. And then the four or so is going to copy and paste all of the data from this first row for the next 364 rows, and then repeat that for the next one. So it's like a loop. Almost. So what we should see in this head is all of the rows here should be zero with all of this information. Let's see if it works. Perfect. So you see all at once. And trust me, there's 365. I don't want to hurt your eyes too much. And then next, what we're going to do is we're going to change the date. So we added the rose, we kept the blending cost or the amortization cost same. But now we need to change the date. So it needs to be 10 29, 31. So what we're doing here with this time Delta index is we're just saying, okay, for the cumulative account, we're just going to add a day after each row. So it's 1029 plus 1030 plus one. That's all this is. Say, let's run that. And here you can see it worked. Right. So this goes on now until 10, 28, 20, 22 when the lifetime of the reservation is all done. So all that's left is to export the file to a CSV on my desktop, and then all we have to do is open it and just make sure it works. So if we look here, we'll see a lot of zero. We'll see the dates keep going up past the five rows that I showed, and the cost is all one dollars. Now, again, as I mentioned, there are 864 of these. So if you go down to the bottom of this, there's almost 315,000 rules. So imagine joining a job and 315,000. I'm very thankful for Python, but this is a project, right? So the product owner was happy. They were glad I didn't spend a week trying to figure this out, and they were able to use it for the book close processes until we found a better automated solution with our standard reporting. So that was a little bit about that example. So now if we go into another example, this one, I want to go over with you. It's a little bit simpler, but it still had a pretty big impact on the analyst's day to day work life. Who I did this for. So the business case here was to concatenate two columns to create a unique Identifier for a financial analyst. They needed this because they needed the unique Identifiers to properly allocate or assign or charge the account to incur the spent for the month. And you see here is a telephone account, and that sounds really simple. Right? All we need to do is take this number and combine it with the date. Excel can do that. No problem. The only issue is there's almost 88,000 rows of data and 47 columns of data. So even though we're not using all the columns, it's still based up the size. And the original solution was just to do it in Excel, right. And equal. You select the number and then the date. But it crashed Excel every single time. So this analyst would pound her head against the wall trying to get it to work each month, and it would crash every hour. So you have to start again and hope that it stays partway through. It was rough. So needless to say, we needed an automated solution here. So what we did is we wrote another Python script with this one. We ingested the file, added a new column formatted, some of the other columns, and then we were able to concatenate it to their business requirements. And I did all of this using Panda. It's relatively simple. And as a result of this effort, the processing time was from more than 30 minutes. I'd say probably more than a couple of hours to less than two minutes. And that means that this analyst was able to do more work in the same amount of time and probably be much happier at the end of that amount of time. So again, that's great. I gave them what I needed. But how could we have gone further? One option is to create a script that reaches for the data each month, performs this calculation, and then spits out a report for her. Another is to make sure that they're strong data quality. Right. Make sure that there's no double unique Identifiers by any chance that'd be bad. Right. So those are some other ways that we could have pushed it a little bit further, but for what they needed, what we did was good enough. So for the demo, I'm not going to do it live. I screenshot only because the file was so massive, and I feel like it's easier to serve talk through. All right. Let me get my highlighter. All right. So I'm going to step through each step here, has the code in one box, and then it has the output, just the first sign of output. I know the head here if you're technical says, like five or three. Right. But I just check first and foremost, we're going to import Panda. That was the package I used and still my favorite. And then we're going to use Pandas to read the Excel file. So if you remember last time it was a CSV, this is an Excel. Pandas is really good at taking almost any Excel version, taking it and being able to read it, which is why I love it. And then all I did here was to make sure that the data was okay. So here we have the closed months we have how much it costs, the invoice number, the invoice date, and then the CK stub, which is sort of like, what is this? And what we really care about, like I mentioned, is this account number and the invoice date. That's what we're going to use. So first and foremost, we need to get this number out from this text, right? We want everything right of this line. So that's what this is doing. I created a variable called Concatenation, and all it does is it takes every value in that CK sub column, and then it splits it wherever. There's this dash line for every single row. And then what I did is I took that variable and I created a new column. So you'll see now from the top to the bottom, the only thing that's changed. I've added the account number with that eight, four, six number in it next to go to the next page. What I did is I changed the date. This is all day formatting based on the business requirements that the analysts wanted. So here we changed the date to a string so that we could play with it. And then we changed it from Dashes. We took out the dash, and then we replaced it with a slash. Right. So here you see 2020, ten one. That's the only change. And then we reorganized it. The analyst wants to be month to eight year. I said, you got it. That's all this is doing is just sort of reorganizing it. And you see us ten, 120 20 rather. And now for the penalty, the big finale. All we're going to do is we're going to create a new column in the data frame called Identifier. And in that Identifier, what it's going to be is the invoice date, which is nice and pretty now, thanks to our previous logic. Plus the account number. What was right of the CK stub as a string? That's it. And then all that was you can see it here. Ten one, 2020, right. And then the account number over here so we can see that it works. And so all that was left is to export it to Excel again, easily doable with Candice and then send it to the analysts to make sure that the requirements and it is they're really happy. And they used it for the next couple of months. So it was a really exciting project. I'm really happy with the output. So that sort of wrapped up my demos of using Python here at Liberty. But I know a lot of people here use it in a bunch of different ways. So, for example, a software engineer who works on our Cloud Sinops team told me he uses Python to pull data from AWS, one of the cloud providers, and perform ad hoc analysis wherever we meet. Additionally, we have a team that creates and maintains our cloud cost transparency reporting to help bringing bring in a lot of data. There's a lot of data in the cloud and ingest it and provide it in new ways for stakeholders that they can better understand their cost. To do that, they use a lot of Python. Additionally, as Stephanie mentioned earlier, you can also create a scraping tool and do some job searching. So there's a lot of endless opportunities. And again, I only shared 3000 330,000 packages that are possible to use. So that's all great. But how can use Python in a nonfinance centered way? Although I love using other potential uses include gathering key of performance indicator metrics for business leadership. So depending on the team or what you're working on in the organization. This can range from gathering sales records, compiling financial results, learning about and tracking employee sentiment. You're involved with the marketing team, you would analyze social media. You can see how many likes do the last TikTok get or how many people commented on the last Instagram post, etc. Etc. Additionally, you can look at it, support tickets and see how long they take to get resolved, which I find important as an end user. If I have an issue with my computer, I want it done in 20 minutes and not two weeks. Ideally. Right. So that's really important. And then finally, compliance data often overlooks, but very important, I've learned, are you complying with accounting, legal and security standards, which are all very important. Another potential use would be to analyze the frequency of delays for customer orders. So think if I'm an Amazon customer, I want it here in two days. I don't want it here in a month. Right. So given all of the supply chain uncertainty right now, this is a really hot topic. It's tracking delays or supply chain issues around the world. And so Python. To sum it all up, I know I just threw a lot at you, but to sum it all up, I love Python because you can use it with almost anything that requires pulling data for multiple systems, compiling it, analyzing it. The opportunities are endless now just about wraps up the webinar. I want to first and foremost, thank you for coming and listening. And I hope now at the end of this, you have a better understanding of how business and technology can work together to provide more efficient business and finance solutions and save you time and money. If you're interested in learning more and applying Python to your career, there are some helpful links at the end of the presentation that we can share with you. I know personally, I found Code Academy really helpful when I was just starting and learning by myself. Also, stack overflow is a fantastic resource. That's again, almost 1.8 million questions. Definitely something to check out. And just to wrap it up, I really believe that as business processes rely more on data, this integration between business and technology is going to be the future of work. Best of luck to you and your future endeavors. And I do want to give a huge thank you to women who code for having me here. It's been a pleasure compiling and presenting this webinar and thank you to everyone on this slide. They helped me endlessly and preferring and preparing and encouraging throughout the process. And now I'll call you for some questions, so feel free to drop them in the chat and we'll try and answer them. Thanks again. Fantastic. So we did have a few questions pop into the Q and A. So can we use Pandas when Excel file has data in multiple sheets? Yes, that one I learned because I came up with. I was like I had 20 sheets, so I didn't want to deal with it. What you can do with Pandas is you can take it in either sheet by sheet or I think you can identify almost like I'm not actually sure. I think you bring it in and then you have to say the sheet you want. Right? So if it's like, bring in workbook one sheet one or workbook two, sheet two, and then I believe you can merge the two data frame and then cross correlate or say like, you need to identify and merge them or whatever you want. Perfect. And we do have time for more questions. So please feel free to drop them in the Q and A. So, Kate, we did have this question. You might have partially answered it already. Could you recommend any good courses or resources to practice applying Python specifically to business and financial sectors specifically? No, I know you Google. There's tons of stuff. I know data. The fish has a lot of good examples as well as W three schools. I think W three schools is the one that I used in my mis classes back at College and Additionally learned Khan Academy had a ton of examples for all different sorts of things. I can't remember if they had anything finance related, but it's definitely a good base work to start to learn the basics. Awesome. And another question, how long did it take to learn Python to get started? It's tricky. So to get started, I said, honestly, one winter break, I just did Code Academy because I was bored at home. So that's sort of how I got started. And then again, I'm by no means a pro, but to get started, it was really just choosing the course. Right. And then from there, being disciplined about doing the work and learning it. But it's like riding a bike, though. If I don't do it for a month, I can go back and be like, oh, that's right. So yeah, I hope that helps. So that's all the questions I have at the moment in the Q and a one in the real working environment. Do you use Jupiter to do all all these the scripting work? Personally, I use Jupiter. I know other people at Liberty use other products. I think I use Jupiter because it doesn't require a license. I believe so. It's really easy and it's free. Right. But I know other people use I'm blanking on the name, but other people use really, whatever suits them best. If not, Jupiter is not the standard by any means, it's probably the minority Google collab. Is that the one you're referencing? I think so. I'm not sure. I think we have time for maybe one more question if anyone has an additional question. Okay. Are there solid packaging? Sorry. Solid packages that will allow you to visualize data? There are. I haven't fully experimented with them. Most of the visualization that I learned in school is with our. But I know that for Python, specifically, there are packages where you can make a graph, right. So you can show a sales trend like I was talking about earlier in Tinter. Like you can add a shape. I'm sure. And I know that there are factors that do graphs and other visuals. I just can't name them off the top of my head. I'm sorry, but I know they're out there. Okay. Fantastic. So I just wanted to end the session with a big thank you to Kate for this wonderful presentation. It was absolutely amazing. I learned a lot, and I'm sure that everyone here learned so much, and we're all going to start implementing these in our real world work experiences. And also a special thank you to Liberty Mutual again for sponsoring this event. And so glad that everyone could join us today. And as you exit the Zoom here today, there should be a feedback survey. If you could please take a moment to complete that, that would be fantastic. All right. Well, thank you so much, Kate, for this. This is just so fantastic. I'm so glad that you joined us today, and it was just so great. Awesome. All right. So thank you, everyone. I'm going to go ahead and close out the session. She loves to be a bridge between finance and technology and has degrees in both finance and mis. When she is not traveling around the world, you can catch her editing academic papers and eating chocolate ice cream. That's wonderful. So without further Ado, I will now pass the session on to Kate. Awesome. Thank you, Stephanie. Hi, everyone. Welcome. Bear with me while I share the presentation. Always the hardest part. All right, let's see. All right. Awesome. So, Hi, everyone. Thank you, Stephanie, again and welcome. I'm super excited to be here today to talk about how we can leverage simple Python process. Simple Python right to improve business and financial processes in your organization and talk about how I've done it here at Liberty Mutual. I'm sure if you've ever worked with Excel, you may have noticed it's not very good at handling the excessive data and tends to crash quite often. But by using coding languages like Python, you can work to avoid crashes and limit human error. The key to this happening, though, lies in the creation of a bridge between business and technology. Personally, I believe that this is the future of work in the industry, and it will lead to more automation of routine tasks and increased analysis being done by humans. I hope that by the end of this webinar, you'll have a good sense of how you can apply Python to get ahead in your career and provide more value to your organization. So let's get started. Here's a quick overview of what I attempt to cover today in this webinar. We're going to start with a brief introduction to Liberty Mutual and myself. We'll go over why we should build a bridge between business and technology. Then we'll look at using Python and business and some good places to start. If it's your first time, then I'll go over examples of how I've used Python here at Liberty Mutual to create efficiencies and show two demos of those in action. Finally, we'll discuss how you can take what we've talked about here today and apply it to your career. Let's get to it. Liberty Mutual is currently the 6th largest property and casualty insurer in the world. Technology underpins everything we do and data plays a huge role in ensuring that all of our customers are taken care of. And that's the reason why we have over 5000 technology employees. We've also been named one of ITG's Insider Pro and Computer Roles Best Places to Work in It for the second consecutive year, we climbed 16 spots since 2020, and we're really, really proud of that. Personally, I'm thankful to be at a place that has been listed here because it's really been helpful in developing my technical skills and knowledge that I hope to share with you today. Over the past year, we've also been recognized by multiple organizations for our continued commitment to our employees. Notably, we've been recognized as one of Forbes Best Employers for Women and for New Grads. Now, both of those I can personally attest to the company being great at it's been really good so far. And then finally, for Liberty here we have a group called Women in Technology that I'm proud to be a part of. I'm currently a member of our recruiting and retention work stream, and we focus on improving the employee and a candidate experience to help attract more women to technology roles and keep them there. We also have other work streams, including Legit, which works to close the gender gap by working with girls to spark their interest in Stem and technology from an early age. I really enjoyed working with Wit so far. And if you want to learn more about what we do, feel free to reach out. If you have any other questions about Liberty or technology at Liberty Mutual, feel free to go to this link to learn more or reach out to me personally. So now a little bit about myself. So I'm based out of Boston, Massachusetts, and I'm currently an associate data analyst for Clouds and Ops Team. My first role out of College, though, is as an associate financial analyst where I worked in the tech finance Department of our company Fun Facts About Me I'm a frequent traveler. The picture on the right is actually taken in Peru in case Alamos on my sweater didn't give it away. And I'm also an avid crossword and puzzle enthusiast. You can catch me at Trivia nights most Wednesdays, and I also try the crossword every day from New York Times. I only do a couple of days a week. Well, but I try, and if you ever want to reach out here some links, they'll be at the end, too. So let's talk a little bit more about my journey and what led me to be talking with you today. I went to College at upstate New York and majored in finance with a minor in mis or management information system, it was there that I began learning how coding could be used in a business context, and even though I am by no means a master coder, right? And we're so dabble in it. I've been really proud of and happy with the impact that it's had so far in my career in such a short amount of time. During my time at RIT, I did a summer internship with Mozilla in San Francisco, where I was a financial analyst intern. This led me to have the opportunity to work with It real estate and Technology Innovation Department to help them manage their spend. It was there that I learned I really wanted to work as an intermediary between finance and technology later in my career. Once I graduated, once I learned this, Liberty Mutual somehow had the perfect job posting. When I was looking financial analysts in technology finance, it was perfect. It was everything I was looking for. And when I started working here, I was placed in the technology Finance Department, which partnered with teams in the enterprise to manage their hosting and server technical spend. However, I was specifically assigned to the Cloud Thin Ops team. This is a group that works across the enterprise to manage spend in the public cloud millions of dollars to spend across AWS, Azure and GCP. Those are the big players. We also create optimization awareness for the entire enterprise, and we help provide cloud education. This means that we're really at the middle between business and technology, right? Because we have to do the financial processes each month to account for the cloud spend. But we also almost act as consultants for the business stakeholders. We partner with them. We help them break down finance terms and technical terms, and we break down the technical terms from the cloud to the finance peers that we have. Right? So we're really right at the intersection. And so you'll notice here that cause Bin Ops has two departments, but I worked with Claudin apps the whole time. That's because earlier this year closing in Ops team moved to the Secure DevOps or SDP Department. This is because they enable developers to easily deploy their apps through pipelines, report on the status of apps in the cloud, and so much more so by moving from technology finance to this organization, we're able to put ourselves at the heart of the developer experience and sort of embed our cost, transparency and effort right where the action is. With this niche positioning, I've been fortunate to have partnered with full time software engineers, architects, business analysts, financial analysts, and so much more in order to get my job done. And as a result of these two roles and partnerships, I've really seen how beneficial new technology can be for improving business processes. In addition to the learning curve engineers have as they learn the business processes. So now I want to sort of share the challenges and the outcomes as a result of building that bridge here at Liberty Mutual. So why build a bridge? Right? Why just swim like you need a bridge? Well, currently, in my opinion, the business and technology Department are oftentimes too siloed in corporate organization, so let's walk through it. Business teams often have standard, routine processes and rely on Excel pretty heavily. And that's not to say Excel isn't great. I love Excel. I come from finance. I live in Excel most of the time, but it definitely comes with its flaws. Right? It can't have a large data and has limited opportunities for automation. Right. Like really good automation. Even though business people may be aware of non Excel solutions, they don't learn how to code in College. And normally Excel is good enough. Right. You can get a good enough to get the job done and be on your Merry way. But they can also feel lost if they say, okay, maybe Excel keeps crashing 20 times in 20 minutes. What's a better way if they want a different way, they often feel lost in knowing where to start looking because they don't learn how to code in College. Right. The technology teams are really great at creating quick, logic based solutions that use code. But the problem is they're a non exposed to business teams and the processes they do very often and B normally they don't learn how to do accrual accounting in College. Right. It's just not part of the curriculum. So there's an almost cognitive dissonance between these departments where there could be really beneficial synergies. So by partnering and working together, efficiencies come created and there's more time for value added work for the organization on both ends. As a result of this, information is more easily accessible and quicker, and as a result, it's easier to make decisions. You can make better informed decisions with more information. So how can we start to build this bridge and bring together both business and technology? A great place to begin is by partnering and creating solutions. Python okay. You may be saying that's great, but why Python? There are so many languages out there, right? Python is great because it's an open source language. This means it's free and there are tons of resources online. So if you're a business person like I am, who wants to become a little bit more technical, Python is great because it has a really easy learning curve and can be a quote, unquote gateway coding language. Right. So once you learn certificate logic, how it works, it can lead to maybe R or C Plus Plus or whatever language you want to jump to next. Right. It can be used to automate a lot of Excel process fees or other standard work. And I think the impact Python can have on Excel is similar to how Excel had an impact on Calculators. Right. Well, you can do a lot of things with Calculators, especially the fancy ones. Nowadays. You can do a lot of things more quickly and efficiently by using Excel. To me, the analogy is Python from Excel, you can do so much more, and it's so powerful. Excel is so good, it still works. But Python is really exponentially better, in my opinion. And since it's open source, that means that people out in the world are actively working on creating their own packages with our sets of tools, features, capabilities to fit their needs. Currently, when I checked earlier this week, there's over 330,000 packages out there. So if you need something to do, I'm sure there's at least one, if not 100, that will suit your needs. And that can be daunting, right. 330,000 packages? Where do I even start? If you ever have questions on writing code or where to start? I found Stack overflow is a great resource. It's an online site that has millions of coding questions and answers. In fact, if you look at the picture on the right, you can see, as of July, there was over 1.8 million questions asked tagged with Python, right. And that's probably just the only ones that were tagged. There's probably more that haven't been. And when I checked earlier this week or earlier today, rather, there's a 6700 questions tagged with Python asked in a week. That's a lot. So if you have a question about Python, the language, the phrasing, the punctuation, someone out there can help you get it done if it hasn't been asked already. And if you look at the bottom right. I know it's really small, but three of the related Tags there are actually packages that I hope to go over with you in a bit more depth. These packages are Pandas, NumPy and Tikinter. These are some of the best packages I've come across in getting started with using Python for business. So let's dive in Pandas. No, it is not a Panda on a computer. Unfortunately, that would be really fun. What it does do is it's a package that is great for beginning to explore the opportunities Python has because you can do so much and it's relatively easy to learn. It's also one of the most popular ones, right? It's not stack overflow. I think it has the most related Tags questions to it. With panels. You can bring in multiple Excel sheet, compare, manipulate, analyze data. You can format data and do some simple machine learning. If you really want to get fancy with it. You can also use this package, for example, to analyze stock trends or something like that, or even emissions data from different countries. It's really popular with finance, statistics and engineering careers. I can attest that. Finance is really popular because it can do so much with Excel. By no means have I done everything Pandas has to offer, but I found it really helpful when I want to automate something from Excel. It's pretty good at doing what I needed to do. Next. We have NumPy. Numpy stands for numerical Python. So this is a great package. If you're looking to do anything computationally, it's really popular in the data science community, and it's really good at performing analysis. It creates and analyzes a race. Think of it as like a list, but it's much quicker than using the standard Python list feature. And you may say, well, why it was so different. It's because it stores the arrays locally and all in one place. Right. So if you have multiple lists, it can quickly look up and calculate and give you that result right away. You can do this for basic algebra statistics. Right. So min Max mode, all that fun stuff, and you can also use it to ensure data quality. Look for missing values or base missing values, et cetera, et cetera. So this one is really good for anything numbers. That's what I like to think of it for. And then Tinter is a package that allows you to create a graphical user interface or GUI and Python. It's built into Python standard library packages. And it's really great if you want to make an interactive report, especially that involves calculations of some kind. Right. So I've seen examples. I remember College. I did an example with Tinter building a sales report. So if you're a salesperson filling out sales order form and you're a manager and you want to see all the sales, it's really popular because it doesn't require a command line interface. In order to interact with it, you can go press buttons, menus, pop up boxes, all that stuff. So it's really easy to streamline and make it easy for the user to understand. So enough about the packages I recommend using. Let's now look at how I use these packages to create Efficiencies during my time of Liberty Mutual. So the first example is one that was given to me in the first month of my career. Here, a finance product owner requested that I am tied for 800 cloud reservation purchases, but let's take a step back. What does that even mean? Amortize 800. What does that mean? So to amortize in finance, it's an accounting principle where you write down the value of an asset over time. So, for example, say you buy a $50,000 car and you expect it to last five years each year you would advertise or write down the value of the car by $10,000. Relatively simple. Except in this example, we're talking about cloud reservation or an upfront payment for one year worth of cloud usage instead of a car, and there's 800 of them. So the product owner wanted to see the cost by day over the lifetime of the reservation. And after getting the assignment, why didn't I just do it in Excel? I tried. Why did I search for an automated solution? Because Excel broke. It was so much data. We had 800 rows and think of it, we have to add 365 rows after each transaction, divide the purchase costs by 365. Copy and paste that 364 times, and then repeat that 800 times more than 800. Actually, it's a lot of work, and I still didn't like how much data and computing was required, especially for such a simple assignment. But it should be a simple assignment, right? Additionally, it would have been a very time consuming TDs and error prone. What if I got tired and I inserted 362 rows instead of 365. How would I know? How would I fix that? There's 800 of them. Right. The risk for little errors like that was really high. So I really liked the idea of coding the work instead of me. So what did I do? So in Python, I wrote a script taking the file. The Excel file. Add 365 rows in between each transaction line, copy the transaction details for the 364 rows, and perform the calculation. For me to do this, I use NumPy from the numbers. Right. And I use Pandas to manipulate the Excel. The outcomes of this for the project that could have taken days or three weeks. I was a bit concerned how long it was going to take. It was done in less than an hour. Right. So huge savings efficiencies. Think of it from the business side. Huge costing, right. Much less compute energy needed less time or more time that I could be doing other things. But that's sort of where I stopped looking at the outcome. I more or less just gave the product owner exactly what they had asked for and didn't do anything else. Now, if you think about it, 800 365. That's like over 300,000 rows. Probably not the best customer service I could have given. Right. So this going further section is really thinking. Okay, how could I have made this better and push the code a little bit further? I could have made a monthly report each month, and I give them exactly what they need to perform their book close analysis. So each month they have to record how much of the reservation was used, which is why they needed it by day. And instead of making them query the data or deal with that massive Excel file, I could create a report each month for them using Python. So that's what I did. Let's now see how I did it. And what I'm going to do is I'm going to share the code and the files. And I just want you to know that all of the data that I'm sharing today is dummy anonymized data, so can't steal our reservation. All right. First, this is the original file I was given the CSV. And what really matters here is the period. So this is the date it was purchased or will be purchased, I guess. And this is the cost. So this is how much we're looking to advertise. For example, it was purchased on October 29, and we need to advertise $365 over $365.$01 a day. Right. Let's see if we can do it. We open up the code what I first did and let me know, Stephanie, if this is too small, but what I first did is important. The Pandas and NumPy packages from Python. Yeah. Could you please Zoom in a bit? You got it. Is that better? Can you Zoom in just a payment? One better. Perfect. Thank you so much. Okay. Thank you. So what we did first is important, the Pandas and NumPy packages that we needed. And then we brought in the file that we need. So this Pandas TD read CSV is bringing in the file. So if we run it and then all this is going to do, we're going to double check that the data looks like it's supposed to remember the Excel file. So we're going to run it and we can see perfect. You see, 1029, 365. Everything else is looking good. The only thing that's really different here is this numbering over here. Right. So instead of the Excel rows, what it is, it created a data frame, and these are just the labels or the numbers or indexes, rather, for each transaction detail or transaction row. So next, what we're going to do is we look at the cost 365 point right now. It's sort of a text. Right. What we need in our two calculations is we need it as a number. So what we're going to do is we're going to convert the blended cost or the upfront cost, and we're going to convert it to a float, which basically is Python for number. So we're going to run that and you can see now the only thing that really changed is the decimal. How many decimals there are. Everything else is the same. So we're golden. Next, we're going to reformat the text of the date. We're going to change it to a date so that we can manipulate it because we need to recognize that over 365 days. So instead of typing in 10, 29, 30, we're going to have this do it for us if we run this or let's see the before. Sorry, 1029. All the slashes. All it's going to do is change it to Dashes and change the data type. So here's Dashes reorganized it. And now Python is able to say that's the date next. What we're going to do is we're going to divide the purchase costs by $365. So the 365 better go to one. Otherwise something is broken. So if we run that, we'll see it worked. Perfect. We can see the one here. The next transaction is also lower and so on and so forth. But the problem is there's only one of them. We need 365 once in order to account for this. Right. What we're going to do is using NumPy. What this code is saying is basically add 365 rows and use a forward fill. So that means we're going to add 365 rows in between each, like zero in between the one, two, so on and so forth for all 864 rows of reservations and then the four. So it's going to copy and paste all of the data from this first row for the next 364 rows and then repeat that for the next one. So it's like a loop. Almost. So what we should see in this head is all of the rows here should be zero with all of this information. Let's see if it works. Perfect. So you see all at once. And trust me, there's $365. I don't want to hurt your eyes too much. And then next, what we're going to do is we're going to change the date. Right. So we added the rose, we kept the blending cost or the amortization cost same. But now we need to change the date. So it needs to be 10, 29, 31. So what we're doing here at this time, Delta index is we're just saying, okay, for the cumulative account, we're just going to add a day after each row. So it's 1029 plus 1030 plus one. That's all this is. Say, let's run that. And here you can see it worked. Right. So this goes on now until 10, 28, 20, 22, when the lifetime of the reservation is all done. So all that's left is to export the file to a CSV on my desktop, and then all we have to do is open it and just make sure it works. So if we look here, we'll see a lot of zero. We'll see the dates keep going up past the five rows that I showed, and the cost is all one dollars. Now, again, as I mentioned, there are 864 of these. So if you go down to the bottom with this, there's almost 315,000 rules. So imagine joining a job and 315,000. I'm very thankful for Python, but this is a project, right? So the product owner was happy. They were glad I didn't spend a week trying to figure this out, and they were able to use it for the close processes until we found a better automated solution with our standard reporting. So that was a little bit about that example. So now if we go into another example, this one, I want to go over with you. It's a little bit simpler, but it still had a pretty big impact on the analyst day to day work life. Who I did this for. So the business case here was to concatenate two columns to create a unique Identifier for a financial analyst. They needed this because they needed the unique Identifier to properly allocate or assign or charge the account to incur the spent for the month. And you see here is a telephone account. And that sounds really simple. Right. Like all we need to do is take this number and combine it with the date. Excel can do that. No problem. The only issue is there's almost 88,000 rows of data and 47 columns of data. So even though we're not using all the columns, it's still based up the size. And the original solution was just to do it in Excel, right. An equal. You select the number and then the date, but it crashed Excel every single time. So this analyst would pound her head against the wall trying to get it to work each month, and it would crash every hour. So you have to start again and hope that it stays partway through. It was rough. So needless to say, we needed an automated solution here. So what we did is we wrote another Python script with this one. We adjusted the file, added a new column formatted, some of the other columns, and then we were able to concatenate it to their business requirements. And I did all of this using Panda. It's relatively simple. And as a result of this effort, the processing time was from more than 30 minutes. I'd say probably more than a couple of hours to less than two minutes. And that means that this analyst was able to do more work in the same amount of time and probably be much happier at the end of that amount of time. So again, that's great. I gave them what I needed. But how could we have gone further? One option is to create a script that reaches further data each month, performs this calculation and then spits out a report for her. Another is to make sure that they're strong data quality. Right. Make sure that there's no double unique Identifiers by any chance that'd be bad. Right. So those are some other ways that we could have pushed it a little bit further, but for what they needed, what we did was good enough. So for the demo, I'm not going to do it live. I screenshot only because the file was so massive, and I feel like it's easier to serve talk through. All right. Let me get my highlighter. All right. So I'm going to step through each step here, has the code in one box, and then it has the output, just the first sign of output. I know the head here if your technical says, like, five or three. Right. But I just have first and foremost, we're going to import Panda. That was the package I used and still my favorite. And then we're going to use Pandas to read the Excel file. So if you remember last time it was a CSV, this is on Excel. Pandas is really good at taking almost any Excel version, taking it and being able to read it, which is why I love it. And then all I did here was to make sure that the data was okay. So here we have the closed month. We have how much it costs, the invoice number, the invoice date, and then the CK stub, which is sort of like, what is this? And what we really care about, like I mentioned, is this account number and the invoice date. That's what we're going to use. So first and foremost, we need to get this number out from this text. Right? We want everything right of this line. So that's what this is doing. I created a variable called Concatenation, and all it does is it takes every value in that CK sub column, and then it splits it wherever. There's this dash line for every single row and then what I did is I took that very well. And I created a new column. So you'll see now from the top to the bottom, the only thing that's changed. I've added the account number with that eight, four, six number in it next to go to the next page. What I did is I changed the date. This is all date formatting based on the business requirements that the analysts wanted. So here we changed the date to a string so that we could play with it. And then we changed it from Dashes. We took out the dash, and then we replaced it with a slash. Right. So here you see, 2020, ten one. That's the only change. And then we reorganized it. The analyst wants to be month to eight year. I said, you got it. That's all this is doing is just sort of reorganizing it. And you can see us ten 120, 20 rather. And now for the penalty, the big finale. All we're going to do is we're going to create a new column in the data frame called Identifier. And in that Identifier, what it's going to be is the invoice date, which is nice and pretty now, thanks to our previous logic. Plus the account number. What was right of the CK stub as a string. That's it. And then all that was, you can see it here. Ten, 120. And then the account number over here so we can see that it works. And so all that was left is to export it to Excel again, easily doable with Candice and then send it to the analysts to make sure that the requirements and it did. They were really happy, and they used it for the next couple of months. So it was a really exciting project. I'm really happy with the output. So that sort of wrapped up my demos of using Python here at Liberty, but I know a lot of people here use it in a bunch of different ways. So, for example, a software engineer who works on our Clouds and Ops team told me he uses Python to pull data from AWS, one of the cloud providers, and perform ad hoc analysis wherever we meet. Additionally, we have a team that creates and maintains our cloud cost transparency reporting to help bringing bring in a lot of data. There's a lot of data in the cloud and ingest it and provide it in new ways for stakeholders that they can better understand their cost. To do that, they use a lot of Python. Additionally, as Stephanie mentioned earlier, you can also create a scraping tool and do some job searching. So there's a lot of endless opportunities. And again, I only shared 3000 330,000 packages that are possible to use. So that's all great. But how can use Python in a nonfinance centered way? Although I love using other potential uses include gathering key of performance indicator metrics for business leadership. So depending on the team or what you're working on in the organization. This can range from gathering sales records, compiling financial results, learning about and tracking employee sentiment. You're involved with the marketing team, you could analyze social media. You can see how many likes do the last ticktocks get or how many people commented on the last Instagram posts, et cetera. Additionally, you can look at it, support tickets, and see how long they take to get resolved, which I find important as an end user. If I have an issue with my computer, I want it done in 20 minutes and not two weeks. Ideally. Right. So that's really important. And then finally, compliance data often overlooks, but very important, I've learned, are you complying with accounting, legal and security standards, which are all very important. Another potential use would be to analyze the frequency of delays for customer orders. So think if I'm an Amazon customer, I want it here in two days. I don't want it here in a month. Right. So given all of the supply chain uncertainty right now, this is a really hot topic. It's tracking delays or supply chain issues around the world. And so Python. To sum it all up, I know I just threw a lot at you, but to sum it all up, I love Python because you can use it with almost anything that requires pulling data for multiple systems, compiling it, analyzing it. The opportunities are endless now just about wraps up the webinar. I want to first and foremost, thank you for coming and listening, and I hope now at the end of this, you have a better understanding of how business and technology can work together to provide more efficient business and finance solutions and save you time and money. If you're interested in learning more and applying Python to your career, there are some helpful links at the end of the presentation that we can share with you. I know personally, I found Code Academy really helpful when I was just starting and learning by myself. Also, stack overflow is a fantastic resource. That's again, almost 1.8 million questions. So definitely something to check out. And just to wrap it up, I really believe that as business processes rely more on data, this integration between business and technology is going to be the future of work. Best of luck to you and your future endeavors. And I do want to give a huge thank you to women who code for having me here. It's been a pleasure compiling and now presenting this webinar and thank you to everyone on this slide. They help me endlessly and preferring and preparing and encouraging throughout the process. And now I'll call here for some questions, so feel free to drop them in the chat and we'll try and answer them. Thanks again. Fantastic. So we did have a few questions pop into the Q and A. So can we use Pandas when Excel file has data in multiple sheets? Yes, that one I learned because I came up with I was like, I had 20 sheets, so I didn't want to deal with it. What you can do with Pandas is you can take it in either sheet by sheet or I think you can identify almost like a I'm not actually sure. I think you bring it in and then you have to say the sheet you want. Right. So if it's like, bring in workbook one sheet one or workbook two, sheet two, and then I believe you can merge the two data frame and then cross correlate or say like, you need to identify and merge them or whatever you want. Perfect. And we do have time for more questions. So please feel free to drop them in the Q and A. So, Kate, we did have this question. You might have partially answered it already. Could you recommend any good courses or resources to practice applying Python specifically to business and financial sectors specifically? No, I know you Google. There's tons of stuff. I know data. The fish has a lot of good examples as well as W three schools. I think W three schools is the one that I used in my mis classes back at College and Additionally learned Khan Academy had a ton of examples for all different sorts of things. I can't remember if they had anything finance related, but it's definitely a good base work to sort of learn the basics. Awesome. And another question, how long does it take to learn Python to get started? It's tricky. So to get started, I say, honestly, one winter break I just did Code Academy because I was bored at home. So that's sort of how I got started. And then again, I'm by no means a pro, but to get started, it was really just choosing the course, right. And then from there being disciplined about doing the work and learning it. But it's like riding a bike, though. If I don't do it for a month, I can go back and be like, oh, that's right. So yeah, I hope that helps. So that's all the questions I have at the moment in the Q and A. There's one in the real working environment. Do you use Jupiter to do all these all the scripting work? Personally, I use Jupiter. I know other people at Liberty use other products. I think I use Jupiter because it doesn't require a license. I believe so. It's really easy and it's free, right. But I know other people use I'm blanking on the name, but other people use really, whatever suits them best. If not, Jupiter is not the standard. By any means, it's probably the minority Google collab. Is that the one you're referencing? I think so. I'm not sure. I think we have time for maybe one more question if anyone has an additional question. Okay. Are there solid packaging? Sorry. Solid packages that will allow you to visualize data? There are. I haven't fully experimented with them. Most of the visualization that I learned in school is with our. But I know that for Python, specifically, there are packages where you can make a graph, right? So you can show a sales trend like I was talking about earlier in Tinter. Like, you can add a shape. I'm sure. And I know that there are factors that do graphs and other visuals. I just can't name them off the top of my head. I'm sorry, but I know they're out there. Okay. Fantastic. So I just wanted to end the session with a big thank you to Kate for this wonderful presentation. It was absolutely amazing. I learned a lot, and I'm sure that everyone here learned so much, and we're all going to start implementing these in our real world work experiences. And also a special thank you to Liberty Mutual again for sponsoring this event. And so glad that everyone could join us today. And as you exit the Zoom here today, there should be a feedback survey. If you could please take a moment to complete that, that would be fantastic. All right. Well, thank you so much, Kate, for this. This was just so fantastic. I'm so glad that you joined us today, and it was just so great and awesome. All right. So thank you, everyone. I'm going to go ahead and close out the site. Session.