OpenBlender is a self-service online platform that enables data scientists to improve the predictive score and quality of Machine Learning models by truly exploiting the immense amount of data available that only a few can benefit from (at a much lesser extent). Users enrich their datasets with correlated variables from their private data repositories and from thousands of live-streamed open sources in no time and without coding. Only a few clicks are needed to join a dataset with others that overlap in time interval or geolocation area. This includes a text vectorizer that blends variables obtained from n-grams (combinations of words) from news articles, social media, emails, call transcripts, chats, etc. to find those that are highly correlated with the user’s target variable in the selected time interval. Visualization tools (maps, timelines, graphs and heatmaps) help users get a better understanding of the data. To end the process, the clean and enriched dataset is easily pulled into a Python or R dataframe with the API. Founded by Antonio Rodríguez Lorenzo, Federico Riveroll, and Javier Echevarría in 2018, OpenBlender is headquartered in San Diego.