When senior Andrea Quiroz was in high school, she thought that math was overrated. That was until she noticed that out of all the subjects, math, which she found the most difficult, just so happened to be the one she got straight A’s in.
“Let’s go to the moon,” her grandfather used to say. “It taught me that I could do anything and be quirky and just be myself,” Quiroz said.
With this conviction to follow her dreams, she declared herself a Business Administration major with an emphasis in Math.
“During student orientation, my dad and I went to the mathematics major booth, and at first, I thought it was kind of weird for a girl,” Quiroz said. “But then, I took my first math class which stimulated my mind in ways my other classes didn’t. Instead of memorizing concepts or expressing my thoughts in papers, I was solving puzzles. That in itself was so satisfying. I even got on the Dean’s List that first semester and so decided to add math to the equation.”
Some of Quiroz’s fondest memories have been studying for her favorite math classes. Currently her happy thoughts include Finance and Microeconomics. From the data she collected in this class, she was able to formulate her senior project.
While researching, Quiroz found the logistics behind how some of the largest corporations like Yelp! bring food-lovers to sushi or taco establishments in town — a project not only inspired by her love of mathematics but also her love for sushi and tacos.
In her paper title “Extracting Key Elements from Yelp! reviews Using Probabilistic & Statistical Methods,” she analyzed Yelp! reviews of restaurants in set locations in order to create recommendations using her own custom code.
“I found out that interesting traits such as ‘a nice manager’ or ‘good garlic bread’ sets it apart from the rest, especially when the reviews had positive tones,” Quiroz said. “After filtering out the data by city and food category, I grouped the reviews and performed an algorithm. As it turned out, the restaurants it recommended were actually the businesses with good star ratings of four to five stars.”
Quiroz used algorithms and statistical measures to analyze the data through her own recommendation generator.
One of the algorithms she used was called term frequency–inverse document frequency (TFIDF), which gave statistical measures on the level of importance for certain words.
“This ruled out common keywords such as ‘a’ or ‘the’ because they’re frequent meaning and not significant compared to other terms that are more definitive like ‘pasta,’” Quiroz explained. Also, by using a Multinomial Learning Algorithm, she was able to load in pre-tagged words that analyzed and deciphered review text as having positive, negative, and neutral words. Quiroz admitted that she did not know math could explain how a search engine lists them in order from positive to negative.
“The algorithms I used helped create probabilities for sentences based on the tone of words,” Quiroz said.
“I think my project brought a new avenue of recommendation that businesses like Yelp! can use to make recommendations more efficiently.”
Working in the field where she gets to manipulate a large data set and apply mathematical techniques is something Quiroz also does outside of class. As a data analyst at HdL Companies, she helps provide sales tax analysis, allocation audits, and tax increment verification for government, city and county-wide agencies.
Some of Quiroz’s future aspirations include becoming a data analyst for other government-run companies — or maybe even Yelp!
“Having both a business- and math-lens, I now realize how underrated math is,” Quiroz said. “It can be applied to virtually everything and anything around me. It never ceases to amaze me. But maybe I’m just a major dork.”