Machine Learning at SQLBits

I had my first look into the data science domain just over a month ago. Here at Village, we had a machine learning research day. The day itself was an interesting experience but was also frustrating - I wanted to know more. So, with a desire to gain more knowledge in the machine learning world, I set out to Manchester to attend the Data Science from scratch with Azure Machine Learning training course at this year’s SQLBits conference.

Comparing with the Village day

The day focused on Azure Databricks – an Apache Spark-based analytics service where machine learning experiments are performed. A number of labs were set out by the instructor to work through the machine learning process. We were looking at a case study about machinery in a factory recording information about errors and failures and the aim was to gain more of an understanding on whether the machinery was going to fail.

Although similar to the Village research day, there was a difference in toolset. For my first experience of machine learning, I used Jupyter Notebook. Although this did the job quite well, Azure Databricks was nice to use, particularly because of its native integration with other Azure services. When it was time to evaluate our results from the machinery case study, we had a brief glimpse into Azure Machine Learning Experiments where the results could be analysed. This was a nice aid to compare how different experiments turned out and it allowed me to visualise what I achieved on the course.

Thoughts on the day

Overall, the trip to SQLBits was an all-round great experience with an interesting insight into the world of data. Since having a few days to digest the conference, I’ve again found that I want to know more about data science and machine learning. Hopefully in the near future there will be some opportunities to apply the knowledge I have learned thus far and expose some solutions to complex data problems.