### Purpose - Create a model or set of models for JoinMyTrip that allows them to increase their conversion from 0.14%. ### Principles - Iterate Rapidly - Create multiple models, keep sharing insights and try to improve constantly. The customer is interested in working with us so we can take our time with this. - Results - We need to help move the needle with results. Even if it means being a consultant if we can move the needle on conversion we can provide serious value and get paid. ### Outcome Visioning We create a few models that are able to predict conversion rates with a high degree of accuracy. JoinMyTrip is able to understand what factors go into conversion and they figure out what changes need to go into their marketing channels, what product features they need to drive and so on. ### Brainstorming - What do I already know about this project? - I have made one model for Preplaced already so I have a decent understanding on how modelling itself works with Azure. I know what basic problems might come up and how the data needs to be structured. - What don’t I know that I need to find out? - I don't exactly know what kind of information comes from Amplitude. - I don't know what attributes and events JoinMyTrip is sending just yet - Not sure how effective our models would be for their use case which is kind of imbalanced. - Who can I talk to who might provide insights? - Once I run a first iteration I can talk to the founder of JoinMyTrip - What can I read or listen to for relevant ideas? - Nothing as of now. Not until I try and get the first round of stuff going. ### Actions - [x] Create a dump of all the events from Amplitude ✅ 2023-07-31 - [x] Make a user list from the API directly if possible or from the events dump ✅ 2023-08-02 - [x] Understand the frequency of events and the nature of attributes ✅ 2023-08-03 - [x] Make a feature list of possible training data ✅ 2023-08-03 - [x] Get numbers for conversion to check with Bharathi ✅ 2023-08-03 - [x] Prepare document for conversation with Bharathi ✅ 2023-08-03 - [x] Mark users as converted and not in training data set ✅ 2023-08-03 - [x] Write a script to generate training data set ✅ 2023-08-04 - [x] Run iterations of model while adding, subtracting and cleaning features as needed ✅ 2023-08-08 - [x] Write a script to generate testing data set ✅ 2023-08-08 - [x] Test the models using a testing data set ✅ 2023-08-09 - [x] Publish results on Bharathi's dashboard ✅ 2023-08-09 - [x] Figure out what to show for Bharathi's graphs ✅ 2023-08-10 - [x] Add actionable insights 📅 2023-08-13 ✅ 2023-08-14 - [x] Upload latest predictions 📅 2023-08-13 ✅ 2023-08-14 - [x] Check if JoinMyTrip is worth pursuing as a user @openos ✅ 2023-09-05