As a business analyst and now data scientist at loeb.nyc
my work spans the analytical spectrum from KPI reporting and dashboard building to predictive models and clustering algorithms.
- For much of my first year I worked with an early-stage startup providing near real-time retail analytics to independent spirits shops: 3x3 Insights. I created all of the user-facing dashboards; helped design and implement the database, and constructed an entire layer of cache tables for reporting purposes.
- Working with skincare company SiO I built software to pull down user information from their email service provider, clean and normalize the data, and create predictions around a user's likelihood to make a purchase.
- At the core of 3x3 Insights value proposition to independent retailers is the ability to see the next big trend in bev-alc coming before the wave has crested. To better inform our partners while preserving anonymity of neighboring stores and competitors, I constructed information-rich clusters of retailers in which to investigate product and pricing movement.
For two years, I was a software engineer at STATS
, the world leader in sports statistics. My work was varied:
- RESTful API
- Many of our largests clients (e.g. Google, Microsoft, Amazon, Apple) hit our APIs with hundreds of calls a minute. I helped maintain and build out new endpoints for these clients while keeping our response times at low latency and live score updates streaming. This is all in c#/.NET
- Stats Requests
- I created statistical reports for, say, the Golden State Warriors, using data from our SportVU tracking system to answer questions like "How directly does Chris Paul drive to the basket vs. Steph Curry?" or "Where on the court is LeBron least and most effective using ball screens?" This involved writing sql queries to calculate statistics from what easily becomes a hundred million rows of data (about one half of the NBA season worth of tracking data).
- Internal Websites
- STATS covers dozens of sports and hundreds of leagues all over the world. A lot of this data is entered by reporters in-venue or watching the game via video, and they need data-entry portals foir us to create a rich statistical report on everything from South African Rugby to English Premier League to International Cricket Council matches. I helped build and maintain these tools, from internal API calls and their oracle procedures to the MVC framework and Angular front-end.
User Experience Contractor
In an effort to increase conversion of trial users to full-time clients, I combed through Kenna Security's user data to generate insights into user behavior. To help marketing and sales teams with user conversion, I created visualizations that demonstrate typical user flows. I made use of Amazon Web Services S3 buckets and API. I also pulled data from SalesForce and Kissmetrics using their APIs to inform the data with qualitative information.
Graduate Student Researcher (M.Sc.)
I completed my graduate work within the Neuroscience and Robotics Group at Northwestern University. My thesis investigated the mechanical variables (forces and torques) associated with rats moving their whiskers against objects. In addition to the biology-side, this work involved computer vision and extensive data QA and maintenance. My findings, based on models of mechanical forces informed by video, and neural recordings from brainstem neurons, were a significant proof of concept.
Quantitative User Experience Research
Using one of the largest datasets Mozilla has ever collected, I quantified user behavior at the per-click level. I created a data-driven story about user interaction with the Firefox browser. I interfaced with members of the business development, security, and design teams to work on UI improvments based on my results. This research is still being used to inform design on the Firefox browser (as of 2016). I was offered a second internship but needed to devote my time to finishing my Master's Degree.
User Experience Research Contractor
I joined the Google [x] User Research team at a pivotal time for Google Glass. The product was slated for shipment to developers in a matter of months and the user interface (UI) needed a fundamental redesign. I ran and aided in numerous participant studies on Project Glass, and a separate comparison of multiple UIs that led to a decision on the primary User Interface and general design improvements. This interface was the one eventually shipped to Explorers a year after I left Google. In my final study, I independently conducted a week-long benchmark comparing performance on Android handheld phones and Project Glass UIs. Core 77 wrote an awesome piece about our group.
In two years at UIC, I was an author on six peer-reviewed publications, generally as the main data analyst. I worked on various studies with the collective goal of establishing neurological biomarkers for pediatric psychological disorders such as ADHD and bipolar disorder. I performed statistical analysis and signal processing on functional magnetic resonance imaging data.