Arilot helps to launch an MVP for Datapost leveraging ML/AI

Datapost is a progressive service for analyzing thousands of data sources like news (NYTimes, sec.gov, SeekingAlpha, etc.), private user storages (Dropbox, Google Drive, etc.). It helps to identify the uptrends, anticipate a decrease of popularity in the media and help to adjust marketing strategies in different regions and even more.

Business Task

We were assigned to Datapost service development leveraging ML/AI. Requested by a user extra-large data from many different sources (news, websites, social media, email, private user storages) should have been processed with Datapost and emotional value (from negative to positive) should also have been detected with it and showed in a user interface.

  • Solutions

    We used Kafka as a Message Bus for communication among all workers of the microservice architecture and store data on GCS. For NLP operations we used spacy.io which is the best Python framework for NLP. Keras and Tensorflow created by Google can help to detect people's emotions. The results of data processing we deposited into a high-performance, scalable timeseries database InfluxDB.

  • Conclusion

    We've launched an MVP successfully and it's being tested at the moment. We continue working on the project until its release.

Other cases

Quick deploy new Transmute environment in few minutes

Blockchain technologies Decentralized Applications