Optimizer for PPC-advertising

My own travel arbitrage startup needed to optimize the cost of contextual advertising (PPC), which requires a mechanism to analyze profitable and unprofitable sales regions. It was required to develop an application that receives data on ticket sales and advertising costs from Yandex Direct as input. The output should contain the ROI of each of the selling cities, the exact maximum cost per click in the specified region, the number of sales and other significant metrics.

To process and map data, the lodash library was used, a parser for the data from Yandex Direct was implemented; Vue.js was used as the front-end framework. The ability to filter by the type of sales (air tickets and train tickets), sales cities (departure and arrival) has been implemented, an algorithm has been implemented for aggregating several transactions into one conversion, calculating the maximum payback price per click in contextual advertising.

The implemented application allowed me to optimize advertising costs, disable ad impressions in unprofitable regions, maximize sales, and identify peak times and days of sales. Thanks to the application, I managed to enter the top-2 webmasters of travel traffic arbitrage.