7Park Data

Alternative Data Analytics Using SQL: Analyzing the current global economic climate

In June, Databricks hosted a webinar with 7Park Data to demonstrate how using SQL on 7Park’s Email Receipt dataset can unlock predictive analytics at scale and help investors understand the current economic climate in near real-time.

Alan Freeman, VP Data Products at 7Park Data (with 10 years of experience in the Hedge Fund Industry in quantitative and fundamental roles) joined Junta Nakai, Global Financial Services Industry Leader at Databricks and Chris Hoshino-Fish, Senior Solutions Architect at Databricks. 

Key takeaways from the webinar include:

  • Alternative data can significantly help augment financial decision-making and can be democratized with the use of SQL 
  • Datasets with a near real-time view into consumer and  business activity, such as 7Park’s Email Receipt dataset, empower investors to make more informed investment decisions and better understand what is going on in financial markets
  • There are tools to accelerate the ingestion and optimization of alternative data to drive predictive analytics and make it easier to get from data to insights

How 7Park Data and Databricks Support Investors

Databricks has a unified data analytics platform that brings together data and AI and capabilities to analyze and build models through one unified platform. 7Park Data is a client of Databricks and our team leverages their tools to explore and analyze unique, new alternative datasets. Databricks tools solve a foundational data problem by making data scalable, reliable and performant and allowing users to leverage new types of datasets for superior decision making. 

7Park Data was founded to help companies make better decisions, and for investors this means providing unique, forward-looking views into company performance and market activity.  

7Park’s data portfolio is vast — ranging from Oncology Treatments (tracking the dispensation of oncology drugs from speciality clinics) to Vehicle Sales (transactions from business & consumer vehicle insurance purchases) to Cloud Infrastructure (actual spend and usage across AWS, GCP, Azure) to Email Receipts (daily transaction data from email receipts) to many more. The Email Receipt dataset is particularly useful during times of economic uncertainty to gain a near real-time view into consumer behavior, and the dataset tracks primarily online purchases which are gaining momentum as people shelter-in-place and minimize trips to physical stores. 

The Email Receipt dataset contains daily transaction data from merchants sharing receipts via email. Within the data, you are able to see customer buying behavior and a view into company performance by tracking order volume, price, basket size and items purchased across merchants. There are over 1.2 million panelists, over 1,100 merchants tracked, over 230 tickers mapped, and history back to October 2012. Additionally, 7Park Data ingests the raw receipt data and normalizes it by building an index to account for shifts in the underlying panel and to reduce bias. 

Email Receipts Shows Consumers Buying more Online Groceries, Pet Supplies, and Home Furnishings

Alan Freeman explores the changes in spending due to COVID-19 observed in the data:

  • In a timeline from February through April, the data shows a massive spike initially in grocery sales (when the WHO declares a pandemic) and a significant change in what people are buying over time as states declare self-isolation and relief checks get deposited in bank accounts.
  • Across the food delivery companies (GrubHub, UberEATS, and Doordash), a surge in gross food sales was captured by the data after the WHO declared a pandemic.
  • The initial hypothesis for Wayfair, an online home furnishings company, was that it would collapse due to people having less disposable income and limited interest in furnishing their homes. Quite the opposite happened. Likely due to everyone spending much of their time at home, furnishing homes (and home offices) surged. Comparing 7Park’s sales index to Wayfair’s stock price, a 100% increase in sales growth in the three weeks leading up to earnings was captured.
  • With item-level detail into what people were buying at Wayfair, one can see a lot of items related to WFH situations were purchased.
  • In early March, the data showed people starting to buy a lot more for their pets on Chewy. Not only were Chewy customers buying more frequently but they were spending more money (captured in an increase in average basket size).
  • Due to the pandemic, rideshare saw a massive drop in bookings. As things start to open up or improve, one could see on a state-by-state basis if there are improvements in rideshare bookings.

With a strong understanding on the type of insights and analysis possible with 7Park’s Email Receipt dataset, Chris Hoshino-Fish, Senior Solutions Architect at Databricks, performs a demo showcasing how one could:

  • Pull the data into Databricks
  • Evaluate what is in the dataset
  • Combine it with other datasets
  • Track historical trends of companies (even before an IPO)
  • Forecast changes over time
  • Perform more advanced analysis to produce different types of forecast (e.g. leveraging ARIMA model)
  • Creating a backtest to validate that the dataset is useful in predicting revenue going forward
  • Identify trends earlier leveraging 7Park Data’s dataset

View the full webinar below and learn how during unprecedented times historical data will not help you try to predict what is happening in the near term. However, with 7Park Data’s daily, granular insights into consumer behavior in near real-time, you’re able to understand where the market is actually heading and how this crisis is actually affecting individual companies.

Alternative Data Analytics Using SQL: Analyzing the current global economic client using alternative data

 

HTML Snippets Powered By : XYZScripts.com