The Art of Data Analysis: Parnikaa's Approach to Solving B2B Production Challenges

Interviewer:

Good afternoon Parnikaa. You just received the best Capstone project award. How does it feel? Could you please introduce yourself?

Parnikaa B R:

Of course, so my name is Parnikaa and currently I'm done with the diploma level. I'm moving into the BSc level soon. First, I would like to express my gratitude to the BDM team for the opportunity to do and share my Capstone project. I would also like to extend a special thanks to Aaditya sir and Ashwin, sir, because they had swift responses to all my queries, and that helped me in the long run. So, I completed my BDM theory in September 2023, and I completed my BDM project in the Jan 2024 semester.?

So, my project is focused on “Enhancing the production Harmony for a business-to-business organization”, which is called Gowthamy stitching unit. So, they face several obstacles. One of them is underlined market demand and labor Supply. So, when there was a high market demand the labor Supply was not available.

Interviewer:

How did you come to know about the BS degree program and data science and what motivated you to take this business problem forward? What were the key problems that you came across that made you decide that this business is something that you were looking to provide solutions to?

Parnikaa B R:

So, the message was circulated through my mom's relative, and I had just completed 12th grade at that point. We checked whether the source was valid, and we saw that it was quite a good opportunity. So, in 2022, I decided to take up the BS degree. And I finished the foundational and diploma level.?

When it came to the BDM project, I was done with the theory. So, I was ready to reach out to a few businesses. And there were a lot of businesses who weren't ready to give the data. They were very willing to help but the data is confidential, so they were quite reluctant. But luckily one of the businesses, Gowthamy Stitching Unit, had data which is also confidential and most of it was handwritten.

So, they gave me a small part of the data which they gave me full permission to use. So even if I had a small amount of data, I was able to find a lot of insights and give solutions to them.?

This project was intensive for me because this was the first time I worked with industry-level data. Negotiating access to the data was also quite challenging because the business data is usually confidential. After successfully Gathering and entering the data, I approached the analysis with a combination of Breadth First Search (BFS) and Depth-first search (DFS) techniques. So first, I would initially do a review, I would see the overview of the data and I would find some irregularities or unusual patterns. Then I would try to go into detail. I would focus on the specifics of that, and I would try to find reasons and solutions for the pattern and throughout the process. I was able to uncover new insights and some of them were also previously unknown to me and the business owner.

And for me the most time-consuming phase was the analysis process because I wanted to get to know the ins and outs of the data. So, it was a long process, but it provided me with a deep understanding of the data and the business operations, which was helpful for my project.?

Interviewer:

Thank you for telling us that. What were your initial thoughts when you enrolled in this program? I’m sure you would have a lot of thoughts in your head. Could you please tell us about it?

Parnikaa B R:

So, in the beginning when I heard that it was online, my expectation was it would be lenient and easy, so I thought maybe I wouldn't have to work too hard, but eventually I was proved wrong because the qualifier was tough. And then I thought “This is serious, I should focus more on it.” Since this was a standalone degree, I decided that this should be my full focus. And the journey has been pretty good so far in the last two years and I have had a good experience with most of the courses. I did find difficulties in some of them, but we had a forum called Discourse. It was extremely helpful. People respond almost immediately and so do the instructors. So, it has been a good experience so far. I'm willing to see how it goes in the future as well.

Interviewer:

That's good, Parnikaa. Very well enunciated from your end. So, moving on to the point of focus, which is about your project. Could you please explain in detail about your project along with the background of the organization and what were the problems that you saw in the business? How did you go with the analysis and the whole process entirely? What are the methods that you used??

Parnikaa B R:

So, the background of my project is a business-to-business (B2B) organization named Gowthamy Stitching Unit. It was started sometime around 2008, and it is associated with a renowned footwear brand. And it's more like a godown for stitching together the basic parts and then they would transport it to the Retail and Wholesale stores. And luckily, I was able to get information from them. So, when it comes to the analysis process, there were a lot of trial-and-error methods, but I found some insights. They didn't seem right. So, I went back to the business owner. I checked with them and some of it was like a man. They were manual mistakes while writing it. After I cleaned all the data since the data was handwritten, I was able to manually fill in the data into Excel sheets.

So, the data cleaning process and data entering process were the same if there was any unusual data. I removed it right then and there. So, after entering the data, I was able to organize the data because data was available but to run an analysis on it. I would need it to be in a properly organized form. So, I did that after that. I thought visualizing the data might help because when I see any UPS or Downs on the graphs, let's say about charts or anything like that. It gives me a slight overview of everything that is going on. So, when I saw the overall trends for three years. The kind of data I got was the revenue. And the employee data because that was the main issue of market demand and labor Supply. So, after I visualized then I noticed any dips which were sudden rises in the data.

I went into the actual details how many employees were there that month? And what was the revenue for that month? And there were also several units coming in the individual art pieces, which were still together. What was the demand and everything? So, I would try to consider a lot of things, most of my insights were intuitive. So, I had a feeling something would be here and then I went into the details, and I verified it. and one of the main issues was the Underlined market demand and? labor supply but there were some other issues that were also coming in as there was a lot of waste within the Unit. There was a lot of cutting and stitching done and there was a lot of waste coming out of that. So maybe there was a way to optimize it which is up to the footwear brand and other issues were covered which affected their business very badly.

Another thing that I noticed was that the labor Supply was quite low because most of their labor originated from North India and the demand was mainly at peak from April to June. This is the time when these employees would usually go back to their natives, which is in North India. Since almost 80% of them were from North India. Most of them would be there in the permanent Workforce. And the temporary Workforce wasn't that effective for them. The issue was that most of their stable permanent workforce was traveling during the time of peak demand. So that is another issue I found. So, the analysis was very iterative, whenever I found something, I would go into detail. I would see what's happening. I would find the reason and then I would find a possible solution for it. After that, I went on to the next one and that took quite a long time.?

Interviewer:

Something like a cause and effect, if I understand correctly.

Parnikaa B R:

Yes, of course. So that was how I did my analysis. I tried to get a complete overview of the data, and I went into all the details as much as possible then finally since we had a report to write I compiled all of them into something understandable. I was able to find solutions, that way I was also able to give some recommendations because right now the world is heading towards a more State of living. Since the leather industry is not particularly the most environmentally friendly, I gave them a few suggestions, you can maybe do something when it comes to ethical sources. Regarding the wastage also, because the leather factory wastage is harmful to the environment. So, I gave them a few suggestions. They said they would take it up with the higher authorities.

Interviewer:

Okay, would you like to please expand more on what kind of analysis you did? What were the unique methods that you employed in your project and the different analysis techniques that you used?

Parnikaa B R:

Of course, mine had a lot of visualizations when it comes to pie charts, box charts, trend lines, etc. I used polynomial transient linear trend lines. I tried to find a pattern between a lot of the data, and I had year-wise data and month-wise data. I tried to do comparisons. And there were a lot of comparisons because I had three years' worth of data. And there was a lot of variation in that. So, I also had a lot of paper tables. Along with a lot of charts and just as an extra step. I decided to employ machine learning. Just to see how the predictions will be the best, not considering any of the trends because that is something for the future. I tried to predict how the industry would look or how their revenue would go for the next five or six years. I believe until 2029.?

I did it and gave basic predictions. I tried to do a test on the 2023 data, and it was quite close with what my machine learning model predicted. So that is also something I recommended to them.

Another thing that I got from my research is that the Indian leather and Footwear industry is going to have an extreme uprising very soon. So, they can take advantage of that and upscale their godown. So, my analysis processes the main part was a Breadth First search and then a Depth-first search in that order. So, I would see an overview of everything and then if I noticed anything unusual or something irregular, I would go deep into the details and I would try to get everything done so that I could get all the insights. My focus was to thoroughly understand the data.?

Interviewer:

Spoken like a true professional in the data science field. It sounds quite interesting when you put it that way. From what I understand, you seem to have a thorough understanding of what your project is, and I guess I believe that's one of the very important skills when you're learning anything based on data science and Analysis, right? So, in that regard could you tell us any of the software that you used and how the learning curve has been so far?

Parnikaa B R:

Overall, from the foundation level to now?

Interviewer:

Yes. So, the foundation level is quite different. You have an introduction to what this is all about. As you progress further, it is different. It's in a way or so more expanded version of the BFS DFS approach that you just said right now. So, as you go deeper and deeper, you understand that there are learning curves, going very differently for you. At what point did you find learning to be quite challenging?

Parnikaa B R:

Okay. So, If I am being quite honest, math. When it comes to numbers, it is something that I find tough. So that is somewhere where I put a lot of focus. I wasn't particularly fond of it, but when I had Math I and Math II, and since MLF and MLT in the diploma levels, they involve a lot of math and if you have a deep understanding, you will do better. And fortunately, I was able to find a lot of good lectures including the ones given by IITM. So, all of that expanded my knowledge of math a lot, especially when it comes to algebra and calculus. And those were the things that I needed, considering it's like a science field. I was trying to put the math part aside. But now I have a lot of knowledge regarding math, and it's helped me in its applications in data science, whether it's machine learning or when it comes to coding.

When it comes to the coding courses as well as Python, Java, and system commands, I had a basic idea of all of this before I entered the degree. I would like to say that before I joined, I knew around 10%. Now, I know almost 70% extra in comparison. So, I did get a lot of information and when they teach you certain things in the courses, they also give you some extra topics that they don't cover in schools. Some of those things I found interesting in them, and I went ahead and researched them as well. The course also piqued my interest in things outside of the syllabus. So, it was very insightful even though they didn't teach it. I sort of got an interest, I should probably go and see how this works, the same interesting thing like that. So, the learning curve has been steep.

I enjoy the process most of the time. I enjoy learning a lot of new things, especially when it comes to machine learning. So, this is something that I have learned to enjoy slowly.?

Interviewer:

That was very nice Parnikaa. From what I can understand you have devoted yourself completely to the whole process. And as you mentioned, you let the process guide you through every step. I believe it is quite an important takeaway for the students and the people who are reading this interview, to ‘Trust the process’. What were the challenges that you faced, and what do you think has helped you overcome them?

Parnikaa B R:

Of course, because in the beginning, I didn't know what to expect and due to COVID issues I wasn't able to write any exams to get into anything. My plan was engineering or something biology-related, but I decided this was a blessing to me because this was the only form of Education I had at that time after 12th grade. So, I took it up and eventually, I decided I could focus on this because this is what I have and I'm grateful for it. I'm still confused as to what might come, but I can take things one at a time. And as of now, it's going well.

However, there were a lot of challenges at the project level because for starters we were told to get the data and perform an analysis. There were a lot of options or options open and the data wasn't given to us people or any specific methods to use so it was all on our intuition. So, first, gathering the data was tough since business data is very confidential. So, I did approach several businesses, and they were encouraging me. But at the same time when it comes to giving their data, they were quite reluctant.

But luckily, I knew someone, a distant relative of mine. His name is Mr. Gaurav, and I'm very thankful for him as the manager of the Gowthamy Stitching Unit. And he agreed to give me the data. It was manually all handwritten data, the employee records, the revenue receipts, everything. They were all printed out there. It wasn't there in any form of a data set on a table or anything. So, everything had to be entered, which took around seven to eight hours. So that was not exactly a challenge. It was just a long process. But the upside of it is that I was able to clean the data right then and there I didn't have any confusion later. So, it was tedious, and I think it was still worth it. It gave me extra ways to understand the data. And the other challenge that I faced was when it came to insights, I was quite confused. I didn't know where exactly to look.

And that's when I realized that once you have the data you must understand the story behind it. You must gather meaning from the data. And that took me quite a lot of time. So, it was a trial-and-error approach. I tried to find out the reasons behind why such a thing happened. Is it a good thing, or how can it be prevented if it's not good? So, then there was a lot of trial and error. The analysis part was the most challenging, but in the end, it was also the most helpful because that's the main part of the BDM project, which is the fact that you must understand the data and find insights.

So, these were the challenges in data collection, data entry, and Analysis. I'd like to stress that the analysis part is the most important in the project because you need to understand the data thoroughly for any part of the project. These were the main challenges for the project. At least when it comes to being a student, I'd like to think that I have it better because there are a lot of discussions that go on in classes as well. Even though it's online or offline, you have a lot of information. Not just within the course, but also on the internet too. There is a lot of information in case there is any confusion, like some YouTube lectures or articles that can help. So that is something that helps as when it comes to knowledge, I didn’t face any challenges in that luckily. But even if I do, there are always people I can reach out to. So, I think that's very convenient for me.

One thing is despite this being mostly online, it is not off campus. You don't have that interaction and those environments. It's still the most you can get from an online degree. So, the challenges are minimal.?

Interviewer:

Thank you for sharing that with us, Parnikaa. This takes us to the last part of the interview. You mentioned the numbers never lie. So, in this regard do you have any words of advice or any message for the students who are aspiring to study the degree course in data science or who are confused as to what to do, when they have enrolled? I believe what you say will help them.

Parnikaa B R:

Of course. I was in the same position and maybe I still am in the same position to some level. But what I would like to advise is regardless of everything else that is going on, this is a piece of general advice, try to work on what you have as of now because once you focus on what you have right now, everything else that you need will come along with each step you take. When it comes to the BS degree to be honest, data science is a multidisciplinary sector. So, you can employ it almost anywhere. If you're willing to keep your mind a bit open when it comes to this, you can even employ it in the medical field as well. Maybe you wanted to become a doctor, but you came this way like me. At least you can employ it in several different fields like Computer science, Healthcare, Environmental field, etc. So, you can try to keep your mind open. As of now, the degree will not give you any specifications. But once you're done with the degree, and once you get a clear idea of what data science is, where it can be used, and how it is helpful, this will help you in the future. Once you get a clear understanding of data science, which this degree is giving. You can see where to go because in the future, of course, the master's degrees. There are a lot of specifications. I have seen a few, they are specific to the environment sciences and there are a lot of different things like that. Whatever you're interested in data science will surely have some or the other place in that kind of field. So, all I would like to say is go ahead with it. Have an open mind, look for opportunities, and at the same time focus on what you have right now. That is my main piece of advice. That's what my parents have also advised me until now and that has been helping me.

Interviewer:

That's very enlightening to listen to you share your thoughts. What I appreciate about taking this interview with you, Parnikaa is that during the entire length of the interview, you have been calm and composed. I guess that is also a lesson that we can take from you. In this regard, this has been an absolute pleasure interviewing you and on behalf of the entire team at the BS degree program in data science, we are proud of you. We commend you for your hard work and efforts. So, congratulations once again!

Parnikaa B R:

Thank you so much. I appreciate this opportunity.