Use of data is held to the same standards of academic integrity as writing. It should be cited properly and used in a way that is honestly representative of what it pertains to. It is important to not falsify or misrepresent data in your work.
The data you use should be as accurate as possible. Being accurate means that it is representative of the truth. For example, if the high temperature at a certain location was 78 degrees Fahrenheit then the number you have in your data set should be 78. While there can be many reasons why some of your data may be inaccurate you should never intentionally change, add, or subtract data for the sake of convenience. It can be tempting to add data when you are short on time. And, when data does not reflect your desired conclusions you might want to change or remove some data. The data collected should drive the conclusions you make, not the other way around.
Data is thought of as a collection of pieces of factual information (facts). Because of its factual nature, rather than it being someone’s expression of their ideas, people often are unsure of when and how to cite data. This relates to the common practice that pieces of common knowledge (information that everyone in your expected audience would already know) does not have to be cited. However, often while the information may be definitely factual in nature, that does not mean everyone knows where the information came from or understands it clearly. Because of that, It is important to cite data. Just as with citing books and other types of sources there are many acceptable citation styles for citing data.
Many major style guides, such as APA, now provide guidelines for citing data. There are also some organizations focused on working with data that have created their own guidelines.
When writing about data make sure you write about reasonable and potential conclusions. One form of honesty is to make sure you are including any relevant information you are aware of into the conclussions you make. Do not ignore some of the data even if it doesn't fit in with your conclusions. When some data does not match with the conclusions you draw then you must either adjust your conclusions or provide an explanation for why some data are part of an exception.