Context and problematic
In an increasingly competitive industry, our client wishes to analyze the content of social networks, to detect future trends.
Goals
Collect a very large amount of data through social networks.
Grasp changes in target audiences.
Detect new consumer segments.
Anticipate ruptures and changes in habits and behaviors.
Our intervention
2 Data Scientist, 1 Data Engineer
- Implementation of Datalake and data pipeline with automated data collection
- Development of text and image analysis algorithms
- Sentiment analysis and scoring of collected data
- Integration of machine learning algorithms for better prediction of trends
- Development of the visualization application in reactJS and VueJS
Résults
Functional platform with updated information in real time.
Our client’s marketing team uses it on a daily basis!
A V2 with a personalized marketing campaign function is considered.
Technical environment
Python – Pandas – Numpy – Sklearn – Tensorflox – Keras – Pytorch
Cassandra – Hbase – ElasticSearch – Apache Kafka
GCP
ReactJS – VueJS
Git