Aktualności

Seminarium Naukowe IGiG – Mgr Inż. Barbara Hordyniec

15-05-2023

IGiG ma przyjemność zaprosić Państwa na seminarium naukowe, które odbędzie się zdalnie z wykorzystaniem platformy Zoom. Prezentację zatytułowaną "Development of an API to support the decision-making process for investing in online advertising campaigns" przedstawi mgr Inż. Barbara Hordyniec z naszego instytutu.



Seminarium rozpocznie się w czwartek 30 marca 2023 r. o godzinie 09:00 AM (CEST).




Dołącz do spotkania Zoom: ZOOM LINK

Identyfikator spotkania: 811 427 0260

Kod dostępu: igig




Development of an API to support the decision-making process for investing in online advertising campaigns

Mgr inż. Barbara Hordyniec, Instytut Geodezji i Geoinformatyki, UPWr




Abstrakt:

Predicting website traffic is crucial for businesses to optimize their online presence and make informed decisions. This work aimed to develop a tool to optimize investing in online advertising campaigns. The proposed API incorporates machine learning methods to predict Google Analytics traffic on a website by integrating weather data. By leveraging weather information, such as temperature, precipitation, and weather conditions, alongside historical website traffic data, a more accurate prediction model can be developed.

The integration of weather data aims to capture the influence of weather conditions on user behavior and website engagement. For example, certain weather patterns might affect people willingness to go outdoors or engage in specific activities, which could subsequently impact their online browsing habits. By considering these factors, businesses can gain valuable insights into potential fluctuations in website traffic and adjust their strategies accordingly.

To implement this approach, historical website traffic data from Google Analytics and corresponding weather data are collected. Features such as date, time, weather conditions, and previous website traffic statistics are extracted. Machine learning techniques are employed to build a predictive model where non-linear trends are fit with yearly, weekly, and daily seasonality, as well as holiday effects.

The model is trained and validated using the collected dataset, and its performance is evaluated using appropriate metrics, such as mean absolute error (MAE) and root mean square error (RMSE). Through experimentation and analysis, insights are gained into the relationship between weather and website traffic patterns.

The result of this study is an API that optimizes the company marketing process by suggesting how to allocate funds for online advertising campaigns based on the results of forecasts. By incorporating weather data, businesses can better anticipate fluctuations in website traffic and develop targeted strategies to optimize their online presence. This approach can be particularly valuable for industries where weather conditions significantly impact consumer behavior, such as tourism, e-commerce, and outdoor activities.