Analytics Vidhya has launched its new series, ‘Leading With Data,’ where industry leaders share their experiences, career journeys, interesting projects, and more. In the debut episode of the series, Mr. Anand S., CEO and Chief Data Scientist at Gramener, was interviewed by the Founder & CEO of Analytics Vidhya, Mr. Kunal Jain. Anand is a visionary in the field of Data Science and a frequent speaker at DataHack Summit. He is known for the exciting projects he works on and the way he narrates stories with detailed visualizations. In this interview, Anand unveils the magic he brings to his talks, sharing with us a glimpse of his storytelling skills and the thought process behind them. Here are some of the thoughtful insights from the interview.

The Art of Storytelling with Anand S.

Kunal J: Every time I attend your talk, I just walk out awestruck by how you narrate the story and the visualizations. Can you walk us through what goes behind creating these talks and how you approach them?

Anand S: There are usually two kinds of talks that I deliver – one aimed at teaching and the other at learning. The learning talk is where I pick something I don’t know enough about, learn about it & then talk about it. It helps me learn something new within a set deadline. It also helps me find out what material works and what doesn’t. So, the next time I talk, I know what to repeat and what to replace or convert to stories. The second kind of talk is where I aim to teach by assembling what has worked well in many of my earlier talks, picking just the interesting parts, and stitching them together.

Kunal J: I’m presuming a lot of what you do in the talks where you are learning would ultimately go into the teaching talks. So how exactly does the actual creation or the research phase for a learning talk happen?

Anand S: I find purpose-driven learning works better than systematic learning. What I mean by that is if I picked up a book or an API’s site and went through it end to end. I learn less than if I pick up a problem and learn to solve it. I also find that if I pick a problem that I’m interested in, I’m more likely to get a solution than a problem that I’m just making up. So, the first task is to figure out a problem that I want to solve. The second part is understanding the problem and trying to solve it, which is a whole lot easier because I have something that I’m interested in.

The Impact of Storytelling

Kunal J: The stories you share in your talks are very interesting. What got you into that format of presenting? Do you find it to be more impactful?

Anand S: People tend to remember the stories told in a talk more often than the rest of what is said. So, it is more effective to convey the content in a story format. While the content of the story is what matters, I’ve realized it is the way these stories are told that makes them memorable.

For example, when I would tell people that older children tend to score more, they’d say I could have guessed that. So I changed the title to say, ‘Do you think the sun sign affects the marks of a child?’ At which point they began to think, ‘No, that can’t be.’ And the reality is, yes, it does – which makes it a hell of a lot more memorable. When I talk about the network of the process of Characters in the Mahabharata, it’s interesting. But it’s only when I started narrating it as, ‘Let’s find out which Pandava was Draupadi’s favorite,’ that it started becoming more memorable. And therefore, how you tell it clearly has a strong impact.

Now, what is interesting in a talk may not necessarily be content-related. To give you an example, I was exploring a talk on the use of comic characters. It struck me that while there are many interesting things about automating comic characters, one particular story particularly resonated both with me and the audience, which is that, as a kid, I wanted to draw cartoons for Disney. And the reality is I can’t; even if you put a gun to my head, I can’t draw. So I gave up. But because I put together, along with my colleagues, this common generation API called Comicgen, using their drawing skills combined with my programming skills – which happen to be used at Star TV, which got acquired by Fox, which then got acquired by Disney – I ended up drawing for Disney, in a very indirect way.

Now, here’s the thing. This is an interesting story but has absolutely nothing to do with the technology behind or the data behind the talk itself. And that’s what I mean when I say it’s not that the storyline necessarily has to be something that strengthens the content directly. It can be something that augments the content in a different way or is a story around it. So I try these out, and whichever works gets distilled and then gets into the talk.

Talking to the Masses vs. Talking to Clients

Kunal J: How is the storytelling in your public talks different from the talks you have with your clients? Does the process remain the same, or is it a very different setup?

Anand S: It’s 50 percent different. There are clients for whom we do almost exactly the same kind of work. It’s just that it’s for private data. So, when a company comes to us with a problem, I would narrate to them a related real-life problem with their data. It would be presented in some kind of sales storytelling format that’ll stay memorable. And then, we’d convert it into an application that delivers the same story.

But that is only 50 percent of what we do with clients because the other 50 percent is, while in a talk, people expect a finished product. While working with clients, they expect two things that are different. First, to be involved as part of the process. So there’s a lot of working process output that goes along. Second, it is the ability for them to be able to deliver such insights. So often, they’d say, “Give me the raw material for me to be able to play around. I don’t want you to come up with the answer. I want to be able to come up with the answer myself. So when it comes to that, what we do is more tool building or solution building that empowers them to find the story and narrate the story. In that sense, it is different.

Anand S.’ Take on Generative AI

Kunal J: As someone who’s really great at building visualizations and data stories, how have your experiments with generative AI been, and how do you feel about it?

AI lowers the cost of doing certain things, and when something has a lower cost, a lot more things become worth doing.

– Anand S., CEO and Chief Data Scientist, Gramener

Anand S: The last time I felt that the world had changed around me this much was when Google was launched. That is when google.stanford.edu came out. This basically meant that I never needed to learn stuff. I could look up anything. And with LLMs coming in, now I realize I never need to think; I just need to learn how to get things to think for me. It is that transformational.

LLMs, I would say, are making the impossible possible. Let me tell you how. A task that would have taken me 10 hours before will take me only 1 hour now, thanks to LLMs. That saves me 9 hours, which I can put into other things, and that’s the difference between impossible and possible. So now, we get better quality output with lower effort.

How do I see it changing the way I work? By and large, I’ve stopped coding. I just tell LLMs to code for me, and there have been a few applications that I have built that are 100 percent LLM generated. In other words, I did not write a single line of code, and dissipation works beautifully. Generative AI, specifically large language models, has given me the ability to create assistants. I can now create an assistant who will code for me or create ideas for me. I can create an assistant for any of a series of different uses. Now that I want to learn by experimenting, I use it as a self-help coach.

If this is the case, I am only limited by my imagination in the kind of assistants I need. This is particularly a problem for me because I’m the kind of person who does not ask for help. Even if I’m stranded and lost in a desert, I won’t ask for directions. It is the change that I’m trying to get to. In 2000 or 1998, when google.stanford.edu was launched, I decided that I needed to make a transition from knowing to being able to discover. Now, I need to make a transition from thinking to being able to guide thinking. It is going to be powerful.

Highlights of Anand S.’s 2023

Kunal J: I’m a subscriber of your blog, and I get this annual reflection that you send out, listing what you’ve achieved in the past year and what you wish to do this year. When did that start? Also, could you please share some of the top topics you are focusing on this year?

Anand S: Well, I started this about 4 or 5 years ago. The main intent was for me just to keep myself honest about the things that I wanted to do that year and share what I’ve learned in the process. Publicly committing to achieving something keeps me driven, knowing I have to update them about it the next year.

This year, I’ve taken on 3 things publicly and a couple of things privately. One is to run 50 experiments and learn from them. I had done this last year as well, but it started and then fizzled out. Again, this year, it started well and then fizzled out. So now I’m trying to build a habit of learning from experimentation.

Another, which is coming on a little better, is changing my environment to see how it changes me in turn. So I try things like working out of a different place every day, meeting a new person every day, listening to a different genre of music every day, etc. The plan was every month, I would run one of these changes in experiments. And some of these have worked dramatically, while others haven’t. It’s a broad process where I commit to sharing and using that feedback to improve.

The third thing I’ve taken up is calendar integrity, which is basically sticking to what I do. This is something that I used to do pretty well. But last year, I felt I slipped a little bit and took it up this year. But it’s more like a habit; once you’ve done it, you don’t really forget.

Some Parting Advice

Kunal J: Based on your perspective and experience gained over the years, what would be your top advice for someone starting off their career today? Or, in other words, what would you advise your own younger self today?

Prioritize the things that you like first. Simply because you find it easier to stick to.

– Anand S., CEO and Chief Data Scientist, Gramener

Anand S: The two are somewhat different. So, first, let me share what I would tell someone starting in their career. The first piece of advice would be to find and explain what you like. It makes work a lot easier when you’re working on what you like. The second thing is to learn by teaching. If you wish to learn something, the best way to do it is by committing to teach it to somebody. That has 3 benefits: First, it forces you to learn. Second, it improves the quality of learning to the point where you can get somebody else to explain. Third, you will make some friends who owe you a favor.

Now, if I had to advise myself, going back a generation, then I’d probably tell myself not to fuss about what’s good or what’s right. Every single thing has a good and a bad – just evaluate both. I say this because I always have, and still do, find an answer to a question without acknowledging that an answer is just one way of looking at a question.

Key Learnings From the Interview

1. In this interview, Anand explained how he creates a framework for his talks. He shared how he prepares for his talks and the difference in approach to learning talks and teaching talks.
2. Anand spoke about the modular parts and the customizable parts of his talks. He sees his talks as experiments to find out what works and distills them into presentable content and templates over time.
3. In his words, Generative AI is the biggest breakthrough since the launch of Google and it is already helping him optimize his workflows.
4. He prioritizes the things that he likes first, which helps him get things done as he finds those things easier to stick to.

Know More

Those were the highlights from our exclusive interview with Gramener CEO, Mr. Anand S. You can watch the full interview here. Stay tuned to our ‘Leading With Data’ series on the Analytics Vidhya Community Platform for more exclusive interviews.