Using OrientDB to Detect Fraudulent Spending by Government Officials

What if average citizens were able to quickly experiment with public government spending data to determine whether any officials were misusing taxpayer funds?

That’s the question Gabriel Mesquita, a software developer and computer scientist from Brazil, recently set out to answer.

In a post on Hacker Noon, Mesquita explored whether any Brazilian government officials were using their monthly allowances illegally by buying products and services from companies owned by people they know.  

In his experimental attempt to detect fraudulent patterns in spending, he turned to OrientDB, the world’s fastest NoSQL database.

A Data Model That Uncovers Anomalies

After obtaining the public data, Mesquita built a data model that leveraged graph database technology. It’s in Portuguese, but here’s what the model looks like:

Gabriel Mesquita data model leveraging graph database technology

 

Mesquita’s model detects which deputies performed multiple transactions with specific companies, whether those companies donated to the specific deputy’s campaign and whether the deputy has any connections, directly or indirectly, to each company in question.

The results? Seven deputies spent their monthly allowances with companies that supported their campaigns in 2014. Another deputy received a donation from a company and then used taxpayer money three different times to support that company.

None of this behavior is illegal, Mesquita suggests.

But, in support of transparency and to serve as another check and balance on politicians, it’s important that taxpayers know about it.

The Takeaways

Mesquita identified two major takeaways from his research:

  1. The Brazilian Democratic Movement (Partido do Movimento Democrático Brasileiro) uses more money than any other political party.
  2. Politicians spend the most money on travel.

Since data pertaining to friends and relatives of politicians isn’t available in Brazil, Mesquita used “fake data” to flesh out his model.

“To validate the model and the whole process I inserted fake data to simulate the fraud scenarios,” Mesquita writes. “Hopefully, if the Chamber of Deputies has this kind of data, they could use the same process to inspect the deputies’ expenses.”

Because he couldn’t access all of the real-world data needed to truly test his thesis, Mesquita’s exercise was experimental in nature.

Still, he found the right tool in OrientDB.

“OrientDB is a great multi-model DBMS,” Mesquita writes. “Graphs are great [at] exposing relationships,” and OrientDB “is a viable solution to find patterns with open data and to provide transparency for our population.”

For more details on Mesquita’s project, read the full piece on Hacker Noon.

To learn more about the world’s leading multi-model graph database and NoSQL solution, visit https://orientdb.com.

 

Venus Picart
Senior Marketing Director, OrientDB, an SAP Company

Take your enterprise to the next level with OrientDB