Research Methodology
We broke our methodology process up into five distinct categories: Define, Research, Organize, Code, and Design.
Define
We first needed to define what exactly we wanted to research and what we were trying to accomplish. We decided that we wanted to research diversity to inform and educate our audience about the importance of diversity across professional industries as well as the improvement or lack of improvement in diversity efforts across specific professional industries. We chose our industries (High Tech, Higher Education, Legal, & Congress) based off of data availability and our interests. We then had to define diversity under the parameters of those four industries.
​
Research
Throughout our extensive research efforts, we ran into challenges finding publicly available data, as many companies and entities are not required to release diversity reports. Finding available data spanning the past 10 years took some digging, especially within the high tech and higher education realms. According to the metadata, diversity is defined under both racial and gender parameters in all four industries, and gender did not extend beyond the female-male binary.
​
Organize
Once we were able to find an appropriate amount of data across our chosen industries, we had to compile them into individual spreadsheets that could be easily translated into Python. We organized the data sets by year, company, race, and gender. We included racial and gender diversity data on the following:
​
High Tech: Apple, Facebook, Google
Higher Education: Texas Christian University, University of Southern California, University of Georgia
Legal Field: American Bar Association
Congress: House of Representatives, and Senate
​
​
Code
Using previous experience in IPython Notebook, also known as Jupyter Notebook, we were able to utilize Seaborn and Matplotlib to create our line graphs comparing our racial and gender diversity data. The code we used is provided in screenshots below:
​
​
​
​
​
​
​
​
​
​
​
​
​
​
In the above screenshot, we imported programs we needed such as Pandas, Seaborn, and Matplotlib. We then set the seaborn style as “dark grid” to provide the graph with a gray grid in the background. We imported our excel file for the specific data sets, and made sure it imported correctly through .head().
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
Next, we created our actual graph, first assigning the size we wanted it to be (fig_dims) and then creating the actual line graph through Seaborn and assigning the title. We repeated this process for each data set to create several different line graphs visualizing diversity trends.
Design
Lastly, we compiled our graphs and research into a website created from a template on Wix.com, making sure to utilize the F-pattern to ensure maximum scannability for our readers. In order to incorporate thoughtful formatting, we used short paragraphs to improve scannability while also keeping in mind that reading long paragraphs on a screen is exhausting. We also used headings and subheadings to help readers visualize the logic and organization of the web-copy. We additionally included bulleted and numbered lists to help the readers recognize important information.