Amazon Internship:
Expanding Alexa COVID-19 Coverage
Role: Data Ontology Intern
Duration: 12 weeks
My Published Data Visualizations on Alexa
“Alexa, what is the percent of people vaccinated for COVID-19 in France?”
“Alexa, how many people got the COVID-19 vaccine near me?”
“Alexa, how many new deaths from COVID-19 were there in Washington?”
Role and Responsibilities
As a Data Ontologist, I:
Analyzed customer data to prioritize and direct projects, documented findings.
Worked with Kettle Pentaho and SQL to extract, transform, and load (ETL) data from external sources to the Alexa Knowledge database.
Expanded and strengthened the ontological structure of the database.
Created map and line chart visualizations to show data on Alexa screened devices.
Worked with a core team (data ontologists) and cross-functional teams (visual designers and software developers) to produce high-quality deliverables.
As a User Researcher, I:
Evaluated high-traffic visualizations and identified bugs and improvements
Wrote a user study proposal to explore customer preferences on information about new movies.
Conducted a user study using a Qualtrics survey with 216 participants from the Alexa InsightsLab.
Analyzed and reported findings to the Alexa Experiences team.
Project 1: County-Level COVID-19 Vaccination Data
“Alexa, how many people are vaccinated for COVID-19 near me?”
Problem Statement: Customers were served COVID-19 vaccination data at the state level when they asked for the county or city level, or asked “near me”.
Solution: Leverage data from the Center for Disease Control and Prevention (CDC) to add to the Alexa Knowledge database and serve to customers.
My Work:
Leveraged CDC API to get daily updated data for county-level COVID-19 vaccinations.
Extracted, transformed, and loaded (ETL) the data into the Alexa Knowledge database.
Created a new table and schema for this data.
Created new ideas and units of “number of new vaccinations” and “kids vaccinated” in the Alexa Knowledge database, so that the data can be asked about and answered by Alexa.
Created map vizzes for county-level COVID-19 vaccination data and line chart vizzes for all other locations.
Going the Extra Mile:
Add existing COVID-19 vaccination data in the Alexa Knowledge database so that all the data is recorded with a unique location and date (previously, just location was used).
Project 2: Vizzes for New COVID-19 Cases & Deaths
“Alexa, how many people died from COVID-19 in Washington today?”
Problem Statement: Customers were served vizzes about the total number of cases and deaths when they asked about the number of cases and deaths from COVID-19 on a specific day (today, yesterday, etc).
Solution: Create new fields for the number of new COVID-19 cases and deaths, and use these fields to answer temporal questions and make line charts.
My Work:
Extracted, transformed, and loaded (ETL) existing COVID-19 data in the Alexa Knowledge database to add new fields for the number of new cases and new deaths.
Created line chart vizzes for these new variables.
Project 3: User Study on Movie Viz
“Alexa, what movies are playing near me?”
Problem Statement: The movie viz was served 60,000/month so optimization has a large impact on the improving customer experience and engagement.
Solution: Run a user study on Alexa customers to ask their preference of seeing information on IMDb ratings versus movie ratings (G, PG, PG-13, R).
My Work:
Wrote a user study proposal to explore customer preferences on information about new movies.
Conducted a user study using a Qualtrics survey with 216 participants from the Alexa InsightsLab.
Analyzed and reported findings to the Alexa Experiences team (IMDb ratings were preferred).