UH endeavor into data science begins by defining the field

Data science in Houston is a loaded term.

It conjures a scuttled University of Texas System project, attractive skills on a job listing and, perhaps, an endorsement by Mayor Sylvester Turner, calling for collaborations between universities to bring Houston into the burgeoning field.

On Thursday, experts in a quiet, sunlit University of Houston classroom made concrete the vague field by giving examples of data science in action. The campus’s new institute hosted roughly 40 people, who listened to field leaders define the term, show how data science can be used in political science and demonstrate a model of coding for large-scale analysis.

UH announced its center in October, but the summer series beginning this month will be the first public events from the institute, meant to fortify the city’s entry into a field that has attracted significant interest around Texas and nationwide.

“This is our maiden voyage,” said Andrea Prosperetti, a UH mechanical engineering professor who directs UH’s institute.

The University of Houston initially said that certificate programs and potentially graduate degrees would begin as early as fall 2018. And in February, Amr Elnashai, UH’s vice president for research and technology transfer, said the university planned to offer “an extensive set” of master’s degrees to integrate data science with engineering, arts, music and other fields.

The institute has not yet moved to create those programs. The Texas Higher Education Coordinating Board has received no applications from UH for any

data science certificates or degrees since October, a spokeswoman for the board said.

Bridging disciplines

Still, introductory courses this fall and the lecture series this summer will put UH on its way, Prosperetti said. UH will aim to offer an unofficial certificate program — giving those who complete the programs recognition for their knowledge but no state-approved certificate — in the coming semester as well, Prosperetti said.

The goal is still to offer coordinating board-recognized programs, he said, adding that the internal approval process at UH would begin in the fall.

“This is not going to happen overnight,” Prosperetti said, adding that the university has talked to Vanderbilt University data science leaders about mutual interest in the field, as well as companies for potential internships.

UH’s first hurdle may well be teaching Houston what data science is.

Andrey Skripnikov knows he’s a data scientist. But ask him what “data scientist” means, and it’ll take longer to explain. Perhaps it’s easier, he said at UH’s Thursday event, to explain what a data scientist is not.

A data scientist isn’t a data specialist, he said. Data specialists take in data and organize it, using programming chops to clean up the databases to make analysis easier.

“They’re the mechanics to the car you’re trying to drive,” he added.

A data scientist also isn’t a data analyst, he added, who take the cleanedup data from specialists and summarize it with basic statistical concepts and a pretty visualization.

Data scientists, he said, ask good questions. They have intuition and a firm knowledge base to know what calculations to do and how to work through large masses of information. And with experience in industry, data scientists think critically about their results.

“Demand for these jobs is growing all over the place,” he said, adding that data scientists work in information technology, education, science, consulting and financial services, adding the ability to critically analyze large amounts of information to teams in various sectors. “It’s the sauce to any dish, the sauce to any meal — the dish, the meal being your majors.”

That approach — introducing data science courses to majors in various fields — is exactly the strategy UH plans to take for its new institute.

UH does not plan to hire faculty specifically for the institute. Instead, Prosperetti said, the institute will funnel money to various academic departments to pay lecturers’ salaries.

“A lot of the activity will reside in the existing colleges, in the existing departments,” he said. “That is how we will reach the entire university.”

In addition, graduate students will do finalproject internships with companies around Houston to grow bigger research projects.

The institute will receive about $1 million from UH’s budget in the next fiscal year, Prosperetti said, along with $2 million from the Cullen Foundation.

Large Texas push

UH is one of many universities in Texas working to bolster data science offerings.

Rice University said it has hired nine faculty for its own data science initiative. In 2015, officials said the university would spend $43 million on the program, including new faculty and staff. Rice will host a conference on data science in October.

In fall 2014, the University of Texas at Austin rebranded its statistics and scientific computation division to a statistics and data sciences department. It offers more than 100 undergraduate and graduate courses to students who study topics as wide-ranging as nursing, education and liberal arts.

And the University of Houston-Downtown offers a bachelor’s degree in data science, which it says is the first of its kind in Texas.

Broad implications

Not every effort has succeeded. A task force convened by the University of Texas System considered using hundreds of acres of land in Houston for a data science institute that would delve deep into health, energy and education. But political headwinds proved too steep. Pushback from regents and local lawmakers forced Chancellor William McRaven to call off the project, admitting he misstepped in buying the land.

UH announced its own institute in the fall. University officials said they first considered creating the program months before UT’s plans were made public.

Prosperetti said he wants UH to make a name for itself in the field by integrating courses with many majors.

“We will reach the entire university,” he said. “The system should be a synonym for data science.” Partially explaining the field’s rapid growth is its potential to influence various disciplines, and on Thursday, that applicability was on display.

Francisco Cantu, a UH political science professor, explained how he collected and analyzed data from photos of voting tallies to determine the extent of election fraud in Mexico’s 1988 presidential election.

The model Cantu created identified fraudulent tallies by examining the layers of data in each photo. The model could, for example, distinguish when someone made a mistake (changing 100 votes for a candidate to 101 votes) from when someone inflated results by hundreds of votes to manipulate the outcome.

Arnie McAdams, attending the lecture, watched with interest.

McAdams, an accountant who just finished his second year at UH’s law school, heard about the data science lecture via email.

And while he watched, he wondered about possible implications of data science on the justice system, knowing that bias influences judicial decisions. Data science, he realized, may be able to measure the scope of bias.

“If you can better understand that,” he said, “you can make a different justice system.”