Predicting Future Weather Conditions With Geospatial Data Annotation for AI

Geospatial Data Annotation, Video Annotation, Polygon Annotation

Mindy Support’s team helped the client increase the accuracy of their AI meteorological prediction model. We completed more than 1,000 videos containing geospatial data with polygon annotation while working under strict deadlines and without sacrificing on quality. The client was very impressed with our work and decided to expand the project with additional geospatial training data for us to annotate.

The client had a sizable dataset of videos to train their AI weather prediction model. This dataset consisted of more than 1,000 hours of videos with various meteorological conditions. The videos were filmed in 30 frames per second (fps), and each frame of the video needed to be labeled with the exact type of weather event, and polygons needed to be drawn around areas of interest, such as cloud formations. 

Since the dataset was quite large, they were looking for a data annotation provider who could assemble a team of 10 data annotators who could complete the entire workload in 5 months. In addition to this, accuracy was of the utmost importance for the client, and they requested that we achieve a quality score of 98%+.

Mindy Support has successfully actualized several projects for the client in the past, which is why the client felt very comfortable entrusting us to do the annotation work. Over the years, we have proven ourselves as a trustworthy and reliable partner, with outstanding reviews from previous projects, which is why the client awarded their project to us. 

We started off by assembling a team of data annotation professionals who had extensive experience in the meteorology industry. Our time-tested methods of quickly recruiting qualified candidates proved very valuable since we were on a tight deadline. Once the team was assembled, we trained everybody on how to use the client’s tool. Thanks to the experience of each team member, it did not take long for them to master the tool and hit the ground running. 

While the annotation work was still in progress, our QA team made sure that all data annotators were performing their jobs up to the highest quality standards. Thanks to the professionalism of our team members and the dedication of our QA department, we were able to meet the client’s expectations without compromising on quality. The client was very happy with the results and plans to expand the project in the future.

• 10 full-time data annotators 
• 1000+ hours of video annotated 
• 98%+ quality score