The field of data science struggles with the presence and successes of women and people of color. According to Ms. Magazine, research suggests that only 15 percent of data scientists are women and fewer than 3 percent are women of color. Due to the education system failing to attract young girls and women to computer science, math, and other related fields, the number of girls and women leaning toward careers in data science is disproportionate. The leaky pipeline metaphor describes the gender gap in STEM-related careers.
A few solutions to this faulty pipeline include STEM education for women and people of color early in life, providing mentorship programs for women in data science, and developing gender-balanced policies.
One of the issues the lack of diversity in data science brings is racial and gender bias in algorithms. Women and people of color become overlooked depending on who is developing these algorithms. Machine learning is the act of training the computer to make judgments or predictions about the information it processes based on patterns it sees. In an article written by Rebecca Heilweil for Vox Recode, Amazon tried to use artificial intelligence to develop a resume screening tool. Its objective was to make screening resumes easier. The issue with that was that the data collected mainly came from men. In the end, this taught the computer to discriminate against women. Amazon decided not to use this tool for several reasons.
We can start accounting for everyone by hiring people of color and women to take on leadership roles. In addition, companies can start using rich and diverse data when training computers to process data.