Responsibilities: Data Science | Machine Learning

Data Science
@ Twitter

‍background 📖

Predicting Lift

Advertisers on Twitter want to know the investment they are making in advertising results in a positive ROI. For lower funnel conversions (purchases on their website) advertisers need to know if seeing an ad led to a purchase.

👉 code for this project 👈

‍the problem❓

Measurment

Gauging the effectiveness of advertising campaigns across platforms has become increasingly challenging. To address this, Twitter introduced a server-side, 1st party pixel that advertisers can integrate into their websites.

DATA SCIENCE💬

I came up with the following plan

The initial phase involved creating a test dataset segmented by iOS and Android devices and further categorized by advertiser. The causal sequence was:
‍
     1. A user views an ad.
     2. A user clicks on the ad.
     3. The user reaches the website's landing page.
     4. The user navigates to the checkout page.
     5. The user completes a transaction.

I constructed a dataset comparing 'before' and 'after' scenarios and applied ordinary least squares (OLS) regression using Python's statsmodels.api.

Once the model was established, I applied it to Twitter's actual data pulled from BigQuery. I used SQL to extract data that closely mirrored the constructed model, then applied the algorithm to understand the potential percentage lift for advertisers

‍THE RESULTS📋

So what happened?

☝️
By implementing Twitter's first-party pixel, advertisers could see a ~15% boost in lower funnel conversion tracking.

✌️
A ~20% increase in customer adoption

👉 code for this project 👈

solution

Native app for management
and optimization communication
between parents and their toddler nanny

Parents Onboarding

After the parent has logged in through one of his social media profiles the app forwards him to chat with a bot that will take the information about the toddler.

Nanny’s Onboarding

All the information that was received during the parent's onboarding will be displayed to the nanny.

Home Instruction

To decrease unnecessary engagement with explanations to the nanny the app will allow adding AR notes around the house.

Home Instruction

So that the nanny can easily find what she needs.

Nanny’s Interaction

On the nanny's home page she will see all the messages left for her and enter the information about the toddler's agenda.

Parents Interaction

On the parents home page they can see all the info collected during the day in a quick and easy scan.