Making sense of challenging data topics one step at a time.

Category: General

Self Service Analytics Does Not Equal Easy

This is a difficult post to write. From a personal perspective, I really struggle with this topic. In my daily work, I assist organizations with enabling a self service data strategy. Collaborating together, we spend hours working to find the right path that matches the needs and goals of the organization with the realities of the organization. I am a huge advocate for self service analytics and believe in it. Because of this, I want to be ridiculously helpful with everyone I encounter.

Sadly, I run into people on a regular basis that want “a little help” to perform some analysis with Power BI or Excel. The problem with this is that the ask is quite complicated. The reality of my situation is that I can make the ask happen quickly. I can spend an hour or two on a report and it will look beautiful. Having worked with Power BI for almost ten years, I have a pile of experience to make it happen efficiently. But the requestor wants me to help them build it themselves. This takes three to five times longer because I am guiding someone through the process. And I am all for doing this because it helps enable another person to use Power BI or Excel.

While I am all for helping people, I find that their interest in learning about these tools is often non-existent. They have a goal in mind that needs to be completed and they think Power BI or Excel is an easy way out. It is a self service analytics tool, so it should be easy! And yet they find the opposite. Without taking taking the time to understand the basics, they end up lost.

What makes this hard for myself and others is that it cheapens the skills we possess as data professionals. Data is not easy. Personally, I find this approach to be a little insulting. So how do we fix this problem?

Setting Expectations for Self Service Analytics

I started responding to every request I receive with clear expectations around self service analytics. I explain the challenges and skills they need to possess before they embark on such a project. Sometimes this scares them off, but it is reasonable way to find alignment.

I often share the basics they will need to learn to be successful. I start with understanding data modeling. We start by talking about things like building a star schema, using a date table, and managing relationships. Sometimes a flat table from Excel or SharePoint is fine, but other times they want more depth and dimension to their reports. Without this in place, they will never achieve their goal.

Next, we discuss how to build out measures. Like every beginner, including myself at one time, they rely on implicit measures. These measures often provide basic aggregations and get the job done. However, they often ask for something more complex. They need to perform time based analysis or use filtering. They are going to need to learn how to write DAX expressions for explicit measures.

Lastly, we will discuss what report design looks like. Often, I have a theme file or template for the organization available, so this helps with branding. But it is important to address best practices with data visualizations. This avoids confusion and provides a better consumer experience.

Assign Some Homework

With expectations set, it is time to assign some homework. Self service analytics requires some investment. But before helping anyone, I assign some basic training. The concept is simple – if someone does not know the basics, they are never going to understand the advanced topics.

Microsoft Learn is my go to for training others around the basics of Power BI. Because it is free, the only real investment is time. It covers a lot of the basics of the tool and allows me to focus on the more advanced topics. I, just like you, need to protect my time. The more I can push to other resources, the easier it is for me to stay focused on my work.

There are two benefits from using this method. First, I can ask for a badge validating that they completed the training. However, I like that this method helps gauge the interest of someone who is asking for help. The course linked above takes about three hours to complete. If they cannot invest three hours in themselves, then why should you invest three hours in them? The answer to that question is up to you, but it sets expectations for the relationship you will encounter through this engagement.

Create Working Sessions

At this point, if they have completed their homework, I recommend creating working sessions to set boundaries. I start with setting up 45 minute sessions with a clear agenda. Ideally, a weekly working session is best. Naturally, you might need to schedule more frequent sessions based upon the business requirements. It is important that you agree upon the cadence to avoid conflict in the future.

For each working session, start with a recap of the previous session. Review the previous topic and any progress made since you last met. Then, work through the assigned topic for the current week. Make sure it is recorded so it can be referenced afterwards. Lastly, decide what the topic of the next working session. This helps keep future sessions focused and your timeline compact.

I recommend that you continue assigning homework. For example, if you have an upcoming working session that will focus on a particular DAX statement, provide resources ahead of time. If you have a measure that needs to use the SWITCH() function, send them to an article like this one to help them understand the concept ahead of time. You can also assign work to be completed in their report before the next session to streamline your working sessions.

As a rule of thumb, I work to keep sessions scheduled. An ad-hoc working session often ends up being a mess. You also have other deliverables and expectations from your leadership that could get in the way. If someone needs time with me between sessions, I generally ask them to schedule it. I feel like the “Can we hop on a quick call? I have a quick question!” conversation always comes at the most inopportune time. Setting these boundaries goes a long way to ensuring you can maintain a balance in your workday.

Provide Some Learning

It never hurts to provide additional learning and training. There are several sources out there. There are several books and sites out there to help. If you need help with Power Query, you can check out my book as an option to assist. If it is with DAX, I recommend The Definitive Guide to DAX by Marco Russo and Alberto Ferrari. Lastly, you can suggest Now You See It by Stephen Few if they need help with data visualization.

At some point, you they will need to embrace self service analytics on their own. By providing materials, they can be focused on their learning and be able to sustain their own solutions. These resources will provide assistance both during your working sessions and in the future.

If self service analytics is a goal, investment in learning is key. I see customers who continue to struggle with this because they do not invest in training and learning. When you are working with an individual, at some point they will need to invest more in themselves to achieve that goal. If you do not help them head down that path, you will be stuck helping them forever.

Conclusion

I love helping people discover the world of self service analytics! I hope you are too! Just make sure you help others in a sustainable manner. These types of relationships are good to have, but you need to manage them well. Be ridiculously helpful – just take care of yourself as well!

Have you encountered this situation? Do you struggle with people who think self service analytics is easy? If so, tell me about it in the comments below!

Sample Data Sets You Can Use for Trying New Ideas

We all want to hone our skills, but sometimes struggle to find good sample data sets to try out new ideas. Sometimes it is about specific data structure or maybe you want to show off an idea but cannot use production data. A lot of what I share comes from real scenarios I have encountered. To share these tips, I have had to use different data sources over the years.

This week, I want to share some data sets that I find fun and helpful for trying ideas out in Power BI. Some will be easier to use than others. You might even need to have a SQL Server to make them work. Regardless, you should be able to find something you can use.

The Most Basic of All Sample Data Sets

If you are brand new to Power BI, the Contoso Financial Sample workbook is a great place to start. It is a free and easy to use data set for beginners. While it does not a great resource for data modeling, it does serve as a quick and easy model to learn the basics of Power BI.

When I was a Power BI trainer, I liked using this data set for basic DAX calculations as some of the key measures such as cost of goods sold (COGS) where included in the model. I can perform some basic calculations which result in net profit.

Check out the Contoso Financial Sample data set here.

Learn How to Find Insights

Another one of my favorite sample data sets that is easy to use is the Pima Indians Diabetes Database from Kaggle.com. Like some of you, I cringe typing out the name of this data set. If published more recently, it would have likely been given a more culturally sensitive name. However, I use this data set for demonstrating the key influencers visual. I have also used it for predictive modeling with Azure Machine Learning, but that is for another day.

This data set was assembled by the National Institute of Diabetes and Digestive and Kidney Diseases. The purpose of the data set was to perform predictive modeling on diabetes in the Pima community. Identified as a high risk population for Type 2 Diabetes, this data represents the Pima community through years of research. Kaggle provides this data set for free. You just need to sign up for an account to access it.

Kaggle is a great resource for other data sets. There are so many to choose from, it is hard to just pick one. However, you are welcome to peruse their catalogue as you might find something interesting. With a little searching, you will find a data set which you can use to build a report on Settlers of Catan!

Check out the Pima Indians Diabetes data set here.

Simplest of SQL Sample Data Sets

Adventure Works is likely the world’s best know SQL database.. A common data set used for training, it is easy to implement. Experience with SQL Server Management Studio will serve you well as you implement this data set. Microsoft provides clear instructions on restoring the database but I find a little extra know how helps. It is wise to make friends with a database administrator if you don’t have one. Offer to buy them a drink or two at happy hour for their help and you will probably make a new friend out of the experience.

Download the Adventure Works data set here.

Binge The Office While Building Reports

Fans of The Office rejoice! TDMitch created a Dunder Mifflin sales data set from the Northwind Traders data base by Microsoft. Just like Adventure Works, this is a SQL data set. Implementing this data set requires additional effort compared to the Adventure Works database. You must follow instructions and run a few SQL scripts to finalize the setup of this data set.

I recommend this data set for someone who is trying to make something that connects with end users. I also recommend this data set for people who are expanding their transact SQL knowledge.

Check out the Dunder Mifflin data set here.

Simplest of REST API Sample Data Sets

REST APIs are great resources for 3rd party data. They work well but you might find frustration with implementing them. I have used this data set before with my series on the Basics of REST APIs in Power BI. While each API endpoint is unique, you can capture the basics using the Yahoo Finance API.

Offered for free up to 100 calls per day, it is an effortless way to learn the basics with no costs. If you are really into stocks, you might even consider purchasing a paid subscription. Spend some time digging through the endpoints and become comfortable with how you can use APIs with Power BI.

You can review the Yahoo Finance API documentation here.

Big Data Sample Data Sets

Sometimes you want to throw a lot of data to test out a solution. The New York City Taxi data set is a massive trove of data that is free for use. Available as CSVs or APIs, you can choose how you want to access the data. I used it to benchmark refresh speeds between various Azure data sources such as blob, table, data lake, and Azure SQL storage solutions.

The Taxi and Limousine Commission provides quality documentation around the data set. It even provides clear descriptions in the data dictionary, maps of taxi zones, and dimension tables. It even explains the difference between yellow taxi, green taxi, and for hire car services.

Check out the NYC Taxi and Limousine Commission data mart here.

Did Not Find Something To Fit Your Needs?

No fear about that! There are tons of free data sources out there for you to use. My favorite place to go is to data.gov and check out different data sets available from the US Federal Government. You can also search for open data from many states and cities. You might even be able to use it for some of your solutions.

Google also has a data set search that will help you find some samples. Search for different topics such as commodities or labor statistics and see what comes back. My only caution is that not every result will be free. However, if you are looking for something specific, this search will help you find what you data you need.

How about you? What are some of your favorite sample data sets? If you have a good one or used one of these, tell me in the comments below!

Welcome and Hello!

Thank you for visiting! I am excited to share my thoughts and experience around all facets of data. I started this journey because of the frustration I often see others suffer from every time they try to effectively use data. My goal is to share some of the tips and tricks I have picked up along the way with tools like Power BI and Excel to simplify the process.

I also am looking forward to sharing my thoughts on the state of data literacy. It has an impact on the readability of content created and the consumers who view it. Frequently I find horrible examples of data visualization out in the wild. Sadly, I see these examples being produced by organizations who claim to hold themselves to high standards. I want to bring these examples to light, dissect them, and show you how to improve them. Report viewers do not always think about these things. Together, we can change how content creators put together reports that are more accessible by consumers.

So thank you for joining me here and letting me share my stories! I look forward to us both learning together!

Powered by WordPress & Theme by Anders Norén