As I have used Sigma Computing in my day to day job and created a Data Coach course on it, I’ve realized a few things I wish I knew when I was first starting using Sigma. Sigma is a great analytics platform that leverages live data connections to data platforms like Snowflake, but in this blog, I will go over a few common mistakes when using Sigma. And of course, we’ll talk about how we can avoid these mistakes as well.
Common Mistake #1: Isn’t Sigma Just a Simple Tool to View Tables?
This answer could be its own lengthy post, but to be brief: absolutely not. It is important to call out that Sigma truly shines by showing large amounts of data in a familiar view like a table, but it can do much more than simply show tables. Here I am connecting to a live Snowflake connection, and I can query a table that contains over 4.5 million rows of data in 3 seconds flat.
![query details in Sigma Computing](https://i0.wp.com/www.phdata.io/wp-content/uploads/2022/07/How-to-Avoid-Common-Mistakes-Made-by-Sigma-Computing-Beginners-Image-1.jpg)
Sigma can also be used for data validation. For example, if you are doing a Snowflake migration, you can conduct an immediate analysis of your data once your Snowflake instance is up and running.
Sigma can also be used to create beautiful, interactive dashboards that let you make business decisions based on aggregate data but also allows you to investigate at the row level with just a few clicks to understand what is driving the visualizations you see in aggregate.
Here is a dashboard created by my colleague, Katrina Johnson, that shows some of the functionality and power of Sigma. Feel free to access this dashboard here on our phData blog.
Common Mistake #2: My Calculations Disappeared from my Table, What Happened?
So often when people first pick up Sigma, they create a ton of calculations on their table and want to create a visualization using some of their calculations. They go to the top left to “Add Element,” select “Data Viz,” select their Snowflake table… and the calculations aren’t there. This can lead to some frustration, but it only takes a slight mindset shift to understand how to understand where to find the calculations.
In Sigma, you can create Child Elements from Data Elements (tables, pivot tables, data vizzes). Say you bring in your Snowflake table with no calculations, just the data from Snowflake. This pulls in the schema from your database and nothing else.
You can create new columns in this table, and from here, you have three options in this example:
- Create a child element to maintain the calculations in your table.
- Create a new data element, select “In Use” for your data source, and select your table with calculations.
- Create a new data element from a new data source and select your Snowflake table.
In the first two examples, your calculations will persist. You are essentially creating a new data element from a table with added columns so that those columns will be brought into the new child element. In the third example, you are bringing in a table directly from Snowflake. These new calculations do not exist in Snowflake, so the calculations will not persist.
In the screenshot above, you can see the first table is a table I brought in from Snowflake and added a new column titled “< or > 3?” The second table is a table I made using the “Add Child Element” button in the top right corner (it shows a bar chart and a plus sign).
Since this is based on the table with the calculation, we have the calculation in the second table too. The third table is a new table brought in from Snowflake. Based on the explanation above, it makes sense that this table does not have the calculation. You can use Sigma’s data lineage function to keep track of your tables, joins, and source elements.
This is important when you have multiple tables, joins, or unions because it can be tricky to keep things in order. Remember, if you want to have calculations persist, always use a child element from your table with calculations.
Common Mistake #3: I Selected a Field for my Filter, Why isn’t it Working?
This is one I fell victim to many times as I began my journey with Sigma. I create a dashboard with a few vizzes and a table, apply a couple of filters, and when I make my selection… nothing changes. Working with filters can be very powerful in Sigma, but again, you need to consider how they work and what the filter is actually doing.
When you create a filter and apply it to a particular data element, like a table, it essentially adds a WHERE ___ = ___ clause at the end of the machine-generated SQL. Take a look below:
This bit of machine-generated SQL shows that I am doing a SELECT * from the Plugs Electronics sample data set. When I add a filter to filter on a particular order number, the end of the query updates, and we see this:
You can clearly see that, when I added a filter, the SQL was updated to include a where clause. If you add a list value control element and select the source for your filter, Sigma will include the values in your filter (i.e., the source will include the order numbers on your order number filter).
However, you need to apply this filter, or this where clause, to a data element. In Sigma, simply select the “Targets” section of your control element to apply a target after you have selected your source:
Here, I selected my source and then went to the target tab to apply the filter to the appropriate target. Don’t forget this part of the puzzle when applying filters, or else you will be scratching your head wondering why the filter isn’t working.
Closing
As you can see, Sigma is a powerful tool that can do much more than just create tables. Whether you are an executive making decisions or a data analyst creating dashboards-Sigma can be used by people all across your organization. Everyone is susceptible to mistakes when learning something new, so I hope this blog can help you avoid some common pitfalls.
No matter where you are in your Sigma journey, we at phData have the resources and expertise to help. Reach out today if you have any questions or need assistance in your Sigma Computing instance!
FAQs
As discussed, Sigma excels at showing data at the row level. Sigma can be great for looking at order information, SKUs, patient-level data (Sigma is HIPAA compliant), understanding metrics like sales by region and what contributes to each region, and many other use cases.
Sigma offers a hands-on lab on their site which is a great resource for getting started. Additionally, Data Coach offers a practitioner course for getting started with Sigma. This course contains over 50 videos to get started with Sigma as well as understanding best practices, culminating with a capstone project! Check out our blog for more resources on getting started.