FOLLOW

FOLLOW

SHARE

Speaker Snapshot: Akash Mukherjee, Data Products, People Growth At Facebook

We speak data visualization with Akash Mukherjee ahead of his presentation in Boston

26Aug

Ahead of his presentation at the Data Visualization Summit in Boston on September 8 & 9, we spoke to Akash Mukherjee, Data Products, People Growth at Facebook.

Akash builds and manages enterprise-grade data products, that enable senior leaders at Facebook to quickly and accurately learn more about their people. Having worked in both engineering as well as customer-facing roles in high-performing organizations like Amazon, Facebook, Tata and Standard & Poor's, Akash has had the privilege to work closely with engineers, statistical analysts, product managers, engineering managers as well as directors and VPs. This gives Akash a unique perspective of how different people in an organization perceive data differently. With this understanding of 'one size does not fit all', Akash has been continuously honing the skill of effectively communicating data, by putting the audience first.

Do you think that the increased use of data visualizations in everyday life has made it possible to make them more complex than before?

Definitely. We often talk about the term 'tech-savviness' when selling software products. I often use a term of choice, called 'data-savviness'. Whenever you're presenting information with numbers and graphs, there's an audience whom you're trying to either inform or persuade. Your strategy of communicating data for effective comprehension by that audience, largely depends on what I call the 'data-savviness' of the user.

With data visualizations growing everyday, through various mediums, online and offline, a huge population is becoming data savvy. They're becoming data literate. And, this audience is not just analysts or leaders who work at tech companies. These are people in all kinds of professions all around the world, who may or may not be dealing with numbers as a part of their day job.

As the average data savviness grows and the fear of numbers & charts reduces, it allows us to explore more creative ways of communicating even more complex data. It is a constant challenge for data visualization community. And, that's exactly what makes it fun.

What do you think makes an effective visualization?

First, let's define what is effective. A visualization is effective when and only when it succeeds in communicating complex information, often rich and big data, to your audience, so that they understand it, they get convinced by your idea or insights, and finally are motivated to take action on it.

Now, 'communicating complex information to your audience' may obviously sound like a cliche. And, it is. So, how do we achieve that? There is no one right answer. Every data visualization is like a new problem you're trying to solve. Every visualization deserves it's own thought process and problem-solving.

I believe there are five parameters that need to be evaluated, every time when you're trying to answer a question with the most optimal way to communicate data.

1) What question are you trying to answer?

2) How big is your data in terms of volumes as well as richness?

3) What is the data-savviness of your audience?

4) What business domain does your question fall in? Is it HR, Marketing, Finance or something else?

5) What are some pre-attentive processing biases that your specific audience has?

What do you think has been the biggest single development in terms of how we visualize data in the past 5 years?

If I had to pick one thing, the biggest development in the area of data visualization has to be the drastic reduction of time to market for each chart published. Products like Tableau, Microstrategy and Qlikview have made it a lot easy to rapidly prototype and launch a brand new interactive data visualization. And, what's great is these tools have pretty decent defaults. So, even if an analyst exploring data using one of these tools, is not an expert in data visualization, the chances of them creating a horribly ineffective data visualization is low.

This is a great move forward because the older generation of BI tools had uninspiring default charts and hence, the visualization's effectiveness relied largely on the acumen of the data visualization artist.

How do you think data viz will change in the next 5 years?

There are three big changes, we will see in the next five years:

1) disruptive tools with serious differentiations,

2) application in new industry verticals, and

3) cross-pollination of people from different backgrounds entering this field

Data visualization is constantly evolving. The number of analytics tools is exploding. There's a mix of two categories of products in the market. On one hand, you have rapid prototyping tools and on the other, there's a range of robust charting libraries in the market. In the next five years, we will see some of these technologies being married. We will also witness the next big disruption in the data visualization & analytics market, since the days when Tableau and Microstrategy came in and stood out from legacy tools of the 90s. So, I expect the upcoming analytics tools to seriously differentiate themselves in their offerings, instead of just playing catch-up with Tableau and Microstrategy.

In the next five years, I also think data visualization will start putting its strong foothold in different industry verticals like HR, Manufacturing, Healthcare, Education etc. We already have some specialized products for Marketing Analytics, Product Analytics, HR Analytics, etc. We will see more of these.

And finally, the real magic will happen when people from varied backgrounds, ranging from astronomy to fine arts come and join the field data science. Some of them may go through the foundational concepts of data. But, in the end, we, as a community, will gain a lot when there is a cross pollination of ideas from these very different universes.

You can catch Akash's presentation at the Data Visualization Summit in Boston on September 8 & 9. 

Comments

comments powered byDisqus
Data culture small

Read next:

Building A Culture Of Data

i