Data, like other information you use in your paper, is evidence to support your thesis. Therefore, you must evaluate data just as you would any other supporting information, such as a scholarly article, webpage, or document. Think of the 6 question words (who, what, where, when, why, and how) when deciding whether to use the data as evidence in your paper. For example, who produced the data? Are they trustworthy? Was their method of obtaining or creating the data credible?
If you are using summary-level data (a chart or graph from a book, website, or scholarly article) you MUST analyze it carefully before you use it. There is a lot of bad data visualization out there and, if you interpret it incorrectly and then use it to bolster your argument, the rest of your writing becomes suspect as well.
If you are using the micro-level or raw data for your analysis, this is less of an issue and you, as the creator of the data visualization need to make sure you keep these ideas in mind.
Note: Interpretations can differ from person to person. This is part of why interpreting data is tricky.
The chart on the left, below, was used in 2015 by Republican Representative Jason Chaffetz of Utah, during the U.S. House Committee Hearing on Planned Parenthood. Look closely at the two lines, they cross each other but are widely different figures.
The chart on the left is the original chart. The chart on the right is with the data graphed more accurately. Even then, the numbers are so different from the top line to the bottom line that they appear skewed even in the graph on the right. This is bad data visualization.