Portfolio
eCommerce Data Pipelines & Dashboards
This project incorporates several of components from our Dashboards as a Service offering. We build data pipelines and dashboards using best-in-class services such as Fivetran, Google BigQuery, dbt, and Looker. We combine this with strategic frameworks to ensure the data, metrics, and dashboards are properly understood across the business.
In this project, we use data from a fictitious e-commerce company, TheLook, to build the technical and strategic components below. The source data are publicly available from Google.
Project Brief | Google Docs
TheLook | Company Metrics Dashboard | Looker Studio
dbt Integration | Github
Dashborad Design | Figma
Data & Metrics Catalog | Google Sheets
Automated Insights (i.e. Semantic Querying) Report
This notebook allows users to ask written business questions about a fictitious e-commerce company, TheLook. The notebook will respond with an answer, transformed Pandas dataframe, and visualization.
This process is known as semantic querying. Semantic querying is one way we can integrate AI into their analytics process.
Anyone can run this notebook, provided they have a Google Cloud Platform account. Check this link on how to set up Google BigQuery.
Sample Questions
What are our monthly sales and customers? Format the month column as YYYY-MM.
Who are our 50 most profitable customers? How much have each of them spend with us?
In what week did we have the highest sales?
"The Most Python" Report
Automated Trading System for Cryptocurrencies
As part of an old "build to learn" project, I built a platform that ingests, analyzes, and predicts cryptocurrency price movements. Eventually, I made successful automated trades based on the predictions and analysis.
The platform was built using the following tools. See the code in Github.
Tech. Jupyter Notebooks, Docker, Kubeflow, Google BigQuery, Google Storage
Libraries. Python, Bash, Scikit-Learn, Matplotlib, Google oauth2 and APIs, Binance API, and more
The How & Why of Enterprise Analytics
A presentation to The Product School in New York City. I discussed the importance and benefits of building an analytics team in a scaleup. At the time, I was leading Spotify's internal data science consulting team, Data SWAT.