5 Must-Have Tools for any Business Intelligence Analyst
- William Auclair-Joyet
- Aug 2, 2020
- 5 min read
Updated: Mar 28, 2021
Business intelligence analysts have broad responsibilities in today's growing digital economy. Indeed, this very term used to be confined to internet only businesses, the likes of pets.com or amazon. Today however, digital is a part of every business, big or small. As the business intelligence analyst of yesterday looked at digital, so does the analyst of today, but at every level of the business. This is why a business intelligence analyst that is serious about data analytics needs a toolkit that is powerful, versatile and adaptable.
Just like the handyman's most important part is their toolkit, so it is for us analysts.
But how do you build yourself a relevant toolkit that can follow you along your most difficult analyses?
That's what this article is for. A simple head-start for you to worry less about how and focus on the what instead.
Snowflake
Python
Tableau
Kite
JASP
1. Snowflake
For your typical business intelligence analyst, Snowflake will be the data bar from which they can access most company data. The advantage of Snowflake is that it's all on the cloud. Using the SQL coding language, analysts can build relational datasets and run database operations online.
For those yet to be initiated to SQL, it is a coding language mainly used to manage data and organize it in a logical way for you to analyze it later. Most basically, SQL or Structured-Query-Language uses the relationships in the data (built or native) to transform or export the data for analysis. A native relation could be the time of purchase of an item, which would natively be on the bill. A built relationship could be the labels your company has associated with the product line for that item. You would then build a database where you have established the relation between the items and their associated product lines.
The power of Snowflake definitely comes from the fact that it is cloud native. This allows you to create views or tables of data based on specific code and data that can be shared with your colleagues for review or accessed for further analysis.
2. PyCharm
No business intelligence toolkit is complete without Python. Considering the breadth of requests that are typical for an analyst, you need a 'hackable' tool to power your code. For me that's PyCharm. Intimidating at first with its feature heavy interface, you will not be disappointed when you overcome that learning curve.
Of the IDE (interfaced developer environment) I have reviewed, it is so far my favourite next to the Anaconda suite (because of Jupyter Notebooks) for its documentation, power and scaleability. You can do anything in PyCharm, from data analysis to machine learning applications. That is why I chose it as my Python sandbox.
With features like the 'Python Console' you can run your code in iterative windows that capture the various objects your code deals with or produces. That way, when you're doing your code review, or troubleshooting that one metric, you can see step-by-step the objects your code transformed or accessed. You can also easily view your dataset row by row, visualize your graphs, and so much more it deserves its own article.
3. Tableau
Now that you have powerful tools to manage and handle data, next is your visualization tool since we all know: an image is worth a thousand word. This is equally true for data analytics! With so many options available, choosing the right data visualization tool is more a function of your company's application suite and technical level. Some great options are out there, including Microsoft Power BI, Google Data Studio, Looker, or Qlick to name a few. As my company uses Tableau Server, I use Tableau Desktop for most of my visualizations.
The key with any data visualization tool is that it allows you to skip the long process of using python packages to build a visualization app. Don't get me wrong, often times you will end up with a better product using seaborn in Python and other packages of the sort rather than using Tableau. This being said, nothing beats a data visualization tool to make a complete c-level dashboard in less than an hour and Tableau sure makes it easy.
What I like best about Tableau is the ease with which you can start. With most of the interactivity being drag and drop and with functionalities one right click away, anyone with a license can get started making graphs within five minutes. Here comes the but: as you progress to more advanced visuals however, Tableau will require you to get extremely creative if the function is not prebuilt. Still, it is very customizable with features like calculated fields that allow you to run SQL-type functions with the results stored in new fields.
4. Kite
Sometimes you just don't know. Sometimes, that documentation on Github is just not that helpful. Don't worry, it happens to the best of us. Everyone forgot how to build their lambda or simply forgot how to call a parameter within your pandas.DataFrame(). That's where Kite comes in. Imagine a little helper powered by AI following you around when you code and highlighting the relevant documentation about what you are building. You can even allow it to follow your mouse, so that as you hover over code or click specific parts, Kite will show you the relevant documentation with examples in your Kite dashboard.
Many times I have relied on Kite to troubleshoot my functions when I build extractors or specific scoring models. In most cases, Kite becomes most useful when I am dealing with a new package. With its explorer, I can lookup the package and find the most common used functions with documentation at the tip of my fingers. You can also set it up for code completion help. With its AI, Kite can automatically detect and complete your code to make life easier for you when you need that analysis ASAP.
5. JASP
Literally, 'A fresh way to do statistics' is their slogan and it fits them to a tee. Open source and based out of Amsterdam, JASP is that kind of software that gives you hope in the internet. Basically made by geeks, for geeks, the software allows you to reach statistical conclusions much faster by automating most of the steps to get there.
With a user-friendly interface, you can easily accomplish bayesian or frequentist analysis with annotated output. For some, this may not seem like much, but as someone who did these in university with RStudio. I can tell you this would have saved me many hours and half the trouble for my final paper should I had known it existed then.
Conclusion
For most business intelligence analyst, picking your toolkit is a luxury. If you have the chance to do so however, or for for those whose department is new enough or company big enough, then it is something you should make sure to prioritize.
In my experience working from home so far, how much fun I have at work is less about who I work with now and more about what I work with.
So here are again the 5 must-have tools for any business intelligence analyst:
Snowflake
Python
Tableau
Kite
JASP
Comentários