Communicating your results to the management is a key skill as a data scientist. There are many tools on the market which can make your life much easier. One of these tools is Tableau. I will review it regarding its functionality and ease of use. You could ask yourself why you need a visualization tool at all. You got Matplotlib, seaborn, ggplot, plotly and so on. But these hard coding visualizations need time. When you are a perfectionist, you can waste tons of time on text alignments and sizes, running your commands over and over again. However, in a professional environment, you need to do things fast and efficient. Managers want you to generate business insights from their data asap. This is why professional companies like Tableau, SAS and IBM (SPSS Modeler) exist. Their tools are mostly pricy, but they (are supposed to) make you more efficient. What about Tableau now?
- Service and Introduction of Tableau
- Ease of use
Service and Introduction
Tableau’s marketing and service is excellent. They offer a free trial version, free introduction videos and free demos in exchange of your personal information. So they will retarget you with emails and events near your location to get a face-to-face contact. As I was already interested in the tool, it was easy for them to close the business. They were always kind and quick. The after sales process was also very good. Their team contacted me and asked me if I am happy or if I need additional support etc.
Tableau is meant to directly communicate with your data warehouse — which makes it extremely handy. However, Tableau can communicate to nearly every data source. From SQL, non SQL databases, via .csv, excel, txt through to spatial data like (tab/mif) you can establish a connection and visualize your data using Tableau. It is also possible to do joins within the tool using different data sources. Additionally, you can rename columns, or factors etc. The possibilities are tremendous and the types of visualization are manifold. You can start with simple bar charts or colorized text, but you can also generate word clouds or spatial visualizations, even with custom json data. Additional to the simple worksheets, you can create dashboards. Here is where the real strengths are, because you can set many filters and every visualization will respond, you can add hover information etc. Especially these responsive elements will look professional when communicating.
Ease of Use
Basically, Tableau is easy to use. All the functions and visuals are nicely explained, and the recommended visualizations of your data help to find a good starting point. But when it came to multiple data sources and their live or extract functions, I still struggle sometimes. One the other hand, you can very easily do data deep dives and answer arising questions using right away. This makes the workflow steady and efficient. A big drawback, however, is the communication to an audience, which does not have a tableau access. This needs some explanation. Tableau has this self BI service philosophy and wants every of your team or even department member to have access to the tool and therefore the data. So everyone can dig into the data an answer questions. I doubt that this is always a good idea. First of all it is costly to buy licences for everyone and secondly people will interpret the data in a false manner. Having analysts knowing the data and their snares is, at least in my experience, helpful. So why is it hard to communicate your results then? It is hard, because of the dashboards. As said before dashboards can be very powerful, but managing them visually (their placement) on different screen resolutions is just a pain. This is mostly because their is no autoscale function as in other tools and you cannot create templates which frame your visuals identically and give them a similar look. Tableau states that this does not fit to their philosophy as described above. They don’t want their general users to be limited by some templates, which would, however, be extremely helpful to the power users. However, I have already seen some dramatic improvements over the past months in the dashboard usability, in fact they listen to their customers and add new features when demanded.
Review Tableau: Summary
|Direct Database connectivity||Dashboard design and adaptaion|
|Broad range of compatible sources||Dashboards have different resolutions|
|Joins of different kind from different sources||sometimes buggy|
|Quick Data manipulation and adaptation||Live/extraction file saves|
|Not costly for a limited amount of users||Philosophy limits core users|