Snowflake - a very cost effective JSON querying tool.
August 06, 2018
Snowflake - a very cost effective JSON querying tool.
Score 7 out of 10
Vetted Review
Verified User
Overall Satisfaction with Snowflake
Snowflake is currently being used to ingest daily JSON files exported from an analytics package into S3. We use Snowflake for Ad Hoc Queries, aggregating daily KPIs and pushing that data into a SQL Server, as well as creating Tableau extracts for our dashboards. We have also started using it for deeper Machine Learning types of analysis - such as creating predictive models.
- Process Engine control - we can stop/pause/start engines for various tasks.
- Processing speed is adequate unless there are many users on at the same time.
- Web interface is easy and intuitive, like the fact that your queries are automatically saved in tabs.
- Very limited amount of tabs - saved queries, which requires us to store the code somewhere else and re-use existing queries.
- Performance can really be a problem if there are many users on the system at the same time.
- SnowFlake support sometimes can be hard to reach.
- We are able to organize all the various data sources into one place and run Ad hoc queries against them.
- Snowflake being very cost effective is definitely saving us $ over their competitors.
- We are able to track KPIs across various games and compare them to each other, thus knowing where to invest our marketing dollars.
Snowflake is cheaper, but TreasureData has a better interface and much more functionality. Also, TreasureData is able to process a greater number of records more efficiently. TreasureData has the ability to run both Hive and Presto. Also, TreasureData has native APIs for many various products, as well as native SDKs. We are currently using both of them for various things.
Using Snowflake
10 - We have several DBAs from Business Intelligence using Snowflake primarily to ensure the data in SnowFlake has been correctly loaded from our S3 buckets. We have several Data Analysts who run ad hoc queries, do in-depth investigations and create dashboards to be be published on Tableau. We also have several Product Managers who sometimes writer their own queries and use Tableau dashboards data analysts have created for them.
7 - Three Database Administrations from the Business Intelligence department, four Data Analysts from the Data Analytics department. We load the data into Snowlfake every morning from our S3 buckets, so it is important for database administrators to ensure the data gets loaded into SnowFlake correctly. Our data analysts run ad hoc queries, do spot checks and create dashboards to be used by our Product Managers.
- Ad Hoc Analysis of various things that are going on in our games.
- In-Depth model building to improve our users' experience.
- Aggregating data and create dashboards to display our KPIs to management.
- We've been able to leverage the cost of Snowflake to get better deals out of our other vendors.
- Machine Learning model building.
- In-Depth analytics investigation into performance of our games.
- Expand our portfolio of games that use SnowFlake.
- Add data from other marketing sources into SnowFlake to integrate with gaming data.
- Build a warehouse sitting on top of SnowFlake with aggregated data.