OMAHA, Neb. – Gautham Pallapa has some closely-held beliefs when it comes to big data, organizational success, and finding your footing in the data science realm.
Pallapa oversees dozens of developers and engineers across multiple areas in his role as vice president of systems and platforms at West Corporation.
He talked with Hot Spots about his career in programming, engineering, data science, and setting people up for success. Here are some of his guiding principles:
-Data by itself does not tell a story.
-Strategic disruption and experimentation lead to progress
-Failure is worth celebrating.
A passion for programming
Pallapa’s path in computer science began as a young student with a choice between programming and music.
“I picked BASIC and did animations, and I didn’t go to the music class. That’s where I started,” he recalled.
BASIC led to C, then to C++, and eventually, an undergraduate degree in electrical engineering.
His post-graduate studies were mostly with Java, but also included Python, a language he has come to love.
“When I started switching over to Python, the thing that I fell in love with was the ease with which you could code and write and develop,” he says.
“It’s so extensible, it’s easy, it’s quick to learn. There are so many resources online, it’s quick to compile, it can run anywhere on any POSIX machine, you don’t really have to add a lot of things, it’s not that memory intensive. You can hack some really cool code out quickly,” he says.
Pallapa says preparation for a career in data science should include learning one high-level, object-oriented programming language thoroughly, and he says Python is a great choice.
“The reason you probably need to learn a language like Python is so you optimally start moving the data in event streams over to [the platform]. Having a ‘nozzle’ where you push the data to it at high frequencies. That’s where Python really gives you the bang for the buck,” he says.
The difference between data and information
From there, he says it’s important for everyone from beginners to board members to know there is a substantial difference between data — and information, knowledge and wisdom.
“Business decisions usually end up being made on data, which probably takes businesses the wrong way. The right way is actually to do it on information,” he says.
The difference, he says, is context.
“I need to understand what was the context in which this data was generated, why did this event happen, what was the action that triggered it. What was the user doing when that event got triggered,” he explains. “Data by itself does not tell a story. We have to tease the story out of it.”
That’s where Pallapa says data visualization becomes so important.
“Sometimes, we as humans find it hard to articulate the patterns in the data, and that’s why data visualization becomes so powerful,” he says. “Your human brain can look at patterns. We’ve been trained through evolution and history to look for patterns. When you display a graph, the reason it appeals to you isn’t just the pretty colors, it helps you articulate what you felt was in your gut.”
And he has a warning for businesses about key performance indicators:
“KPIs should never be raw data. Multiple pieces of data should roll into a key performance indicator. And I think some companies do it the naive way where they say only these different data elements are what we’re going to analyze,” he says. “The number of clicks is a false positive for a KPI.”
Blurred lines
Pallapa says the lines between data science, business intelligence and statistics can be blurry.
“They tend to play off each other. Data science, in my mind, is actually the entire scientific realm around data and how you treat it with a scientific approach,” he says. “Business intelligence is more around trying to tease out patterns and information in order to trend and forecast, and if you are getting your ROI or whatever your KPIs are for your business.”
Whatever path a student wants to pursue, Pallapa says there are some basics to build along the way — among them, a portfolio and a Git repository.
“I encourage my candidates to show me their repositories and we go through the code. It also shows the hygiene of the code, how you are coding, the thought process you have, so if you are comfortable enough to share and show the code, that means you are confident of your development skills. That comes off in a very different way to a hiring manager,” he says.
The best way to build a portfolio? Find a problem that annoys you.
“And write an application that solves that problem. You don’t have to solve a world problem, it can be your own personal problem,” he says.
As an example, Pallapa talks about an intern he worked with who wanted a better way to handle the logging duties he was assigned. The intern wrote a program to automate some of those tasks, shared it with others and impressed his managers.
‘We celebrate failures here’
That kind of experimentation can have big payoffs, Pallapa says.
“The concept of experimentation is really dear to my heart. Taking huge, turnkey projects and turning them into continuous experiments to meet our targets,” he says. “So for example, we run three or four experiments in the teams and figure out if we’ve succeeded or we’ve failed. And we celebrate failures here.”
Pallapa says experimentation in a safe environment fosters accelerated learning.
“The failure of one person is valuable learning for the rest of the group,” he says.
Finally, Pallapa says it’s important to have fun along the way. As part of that, he helps organize game nights where West employees work in teams to solve brain-teasing problems.
“My primary belief is happy people are productive people. And my second belief is that strategic disruption leads to progress,” he says.
*****
Related:
Follow Gautham Pallapa on LinkedIn
Interface Web School’s Fundamentals of Data Science Certificate