Capstone – Project
Using Neural Networks to Unearth Patterns of Consumer Behavior
A Collaboration Between
Consumer Orbit, is a Consumer Science company that utilizes Big Data to help clients answer marketing and business questions about their customers. We work with clients to find data-driven answers to challenging business questions.
Through our patented approach to data integration, we turn big data into action. We have combined the best available consumer behavior, attitude and media preference data with technology that allows us to create data-driven solutions designed to solve each client’s unique business challenges.
The results are strategic, actionable and measurable, allowing clients to make informed decisions about things like media budget allocation, sales team alignment and product assortments at the market or store level.
For this project Consumer Orbit is offering a non-statistical sample of our time-series online search data looking for more insight into which statistical tools are best at discovering consumption patterns for the American consumer and using those consumption patterns to predict behavior.
Key project milestones:
1. Get Up-To-Speed on Consumer Orbit’s Data & Design the Input and Output of the Neural Network
2. Conduct and carry out initial predictive modeling/classification/regression through the use of LSTM (Long-Short Term Memory) neural network architecture of online search data
3. Reexamine online search data with DNC (Differentiable Neural Computer) architecture to discover and predict consumer behavior patterns
4. Perform a comparative study and theoretical analysis of the two different results
5. Report the results to the company. A special attention will be given to potential bias and application research.
Experiential Learning Program Details
|School||University of the District of Columbia (UDC)|
|Engagement Format||Capstone - Small Team Consulting Project - Students work in small groups of 2-6 directly with faculty and host company project champions on developing real solutions to real-world challenges.|
|Meeting Day & Time||Monday and Wednesday 4:30pm-5:50pm|
|Student Time Commitment||4-7 Hours Per Week|
|Company Time Commitment||1 Hour|
|Touchpoints & Assignments||Due Date||Submission|
Key Project Milestones
February 24, 2020 - KICKOFF MEETING @ 5:00 PM ET
Kristi, the Consumer Orbit’s SVP of Analytics / CIO will web conference into class to discuss her product, vision, and the project scope
Prepare for kickoff meeting by familiarizing yourself with Consumer Orbit’s website: https://www.consumerorbit.com
March 6, 2020 - Get Up-To-Speed on Consumer Orbit's Data
Online search data sample and schema from across 200 different retailer and shopping comparison websites:
- How is the data generated?
- How large is the data set?
- What is the data schema and structure?
- Design the input and output of the neural network
March 27, 2020 - Conduct and carry out initial predictive modeling/classification/regression through the use of LSTM (Long-Short Term Memory) neural network architecture of online search data
- Conduct and carry out initial predictive modeling/classification/regression
- Complete LSTM Portion of the project
April 17, 2020 - Reexamine online search data with DNC (Differentiable Neural Computer) architecture to discover and predict consumer behavior patterns
May 6, 2020 - Perform a comparative study and theoretical analysis of the two different results
Pay special attention will be given to potential bias and application research.
Report the results to the company
There are no resources currently available