The views in this article are my own and do not represent those of any organization with which I am affiliated.
There are a few goals for writing this article:
- Providing recommendations to companies on preventing sexual harassment and effectively managing sexual harassment reports
- Creating a genuinely welcoming and supportive work environment, one that truly embodies diversity and inclusion
- Encouraging and supporting discourse in the community
Part 1: Background
There has been substantial public opinion regarding the latest tech #MeToo scandal which was revealed publicly on April 4, 2019 and what is known as the “DataCamp sexual harassment case.” Some of those articles are listed in this DataCamp Sexual Harassment Timeline of Events. There are three statements issued publicly by DataCamp including an apology by the “offender.” Most recently, the “target” shared her story in a May 13, 2019 Buzzfeed investigative report.
Language is critical to understanding and impartially representing the entities involved. Here are some definitions used in the article:
- Target: “Target” is the preferred term over the word “victim” by the R-Ladies community to reference the woman who experienced and suffered through this incident.
- Offender: Tech company CEO who sexually harassed the target
I could not recall when I first heard of DataCamp (DC). So I searched my inbox and found that it was May 2015 when a speaker for my meetup group Women in Machine Learning & Data Science (WiMLDS) had recommended the free Intro to R course for our attendees in preparation for a workshop I was organizing. About two years later, in March 2017, our meetup group had shared a 1-month free DC discount with our community. In June 2017, by way of recommendation from an R-Ladies member, I had an appointment to meet with DC’s Head of Content to discuss doing some work for DC. But, there was a major subway derailment in NYC that day which hindered my plans and our meeting was canceled. What followed was the typical busyness of New Yorkers, extended summer vacation schedules and emails sliding deeper into the electronic pit of our inboxes. Thus, the meeting never did get rescheduled.
Over the next year, DC was proactively recruiting instructors and their focus on increasing diversity impressed me. One of my friends was also creating a course for them. So, in August of 2018, I submitted an application to create a course for DC. I first created an extensive outline of my course for which I invested 40+ hours. The next step was recording a screencast. After investing significant time into designing and recording the initial screencast (40+ hours), I was asked to resubmit with revisions. All that I learned about my course topic which I was preparing was unregrettable, and yet I decided to abandon the course. Time was a commodity, there were other commitments and risk that I could invest another 40+ hours without a course approval or any compensation. Successfully working with DC proved once again to be elusive.
Meanwhile, my friend was still creating her course and the development had persisted for over a year. She mentioned that each of her contacts was leaving DC and it was difficult to maintain productive communication or progress. We wondered what was happening over at DC. We speculated there was likely some of the typical office politics at work or start-up woes with the employee turnover, but we did not have any specific information.
DataCamp Online Courses
In the Fall of 2017, I was teaching my 8-year-old niece coding. We used a beginner coding book which did not provide as much utility as I expected. My niece does not know how to type, so typing in 10 lines of code was tedious for her. I thought a possible solution would be to teach her to type first and we worked on that during my extended visit at her house.
During my 2018 winter holiday season visit to New Jersey, I once again sat with my niece (then 9 years old) to teach her Python. We explored half a dozen websites and DataCamp proved to be the most user-friendly (no Python installs required, limited typing, good explanations, etc), and she began the Intro to Python course and loved it. I received a free one year subscription through a non-profit, and forwarded it to my niece so she could continue learning.
In January 2018, I was invited to be on a panel discussion to celebrate the launch of the DataFramed podcast. There were 2 men and 2 women on the panel, and it was moderated by Hugo Bowne-Anderson. At the event, David Robinson announced he was leaving Stack Overflow and would be joining DataCamp as Chief Data Scientist. I had met David through his sister Emily Robinson, who I first met in the classroom at a data science bootcamp during the summer of 2016 and where David had been a guest speaker. David had been very supportive of my meetup group, WiMLDS. He had hosted my meetup group at Stack Exchange on a Saturday on 3 separate weekend occasions (January to March 2017), arriving earlier in the morning than us organizers to set up and staying later to finish up (neural networks workshop, scikit sprint, Smart Cities Hack ). There is a dearth of availability of weekend meetup space, and so David hosting our group had been immensely valuable and his support of women in data science was genuine and preceded his work at DataCamp. Additionally he is an ally for women in data science; one example is his recommending me to be a panelist that day. It was an appreciable opportunity, one that apported me and my meetup group positive visibility. As well, the questions under discussion at the panel that day gave momentum to blogs that I would write later that year.
In mid-November 2018, Hugo reached out to me to discuss being interviewed on DataFramed. We had a prep call which ended in a passing question about why the R community was more inclusive than Python. I spent the long Thanksgiving weekend researching that topic and wrote an article on Python / R inclusivity. The podcast was recorded in mid-December, right before the holiday season and released in late February 2019, Women in Data Science. As a result of the global reach of that podcast, WiMLDS increased the number of chapters by 50% in less than 2 months, from 39 to 60+. WiMLDS was flooded with inquiries from around the world on how to connect with the community of women in data science and machine learning.
The April 2019 Breaking Story
Some weeks prior to the April 4, 2019 statement released by DataCamp, I had heard a few passing references about a “sexual assault” incident at DataCamp, but any additional details were limited. With the modicum of information, comments and whispers and certain verbiage, there was misknowledge that a woman had been raped by a founder of the tech company with absolutely no acknowledgment, penalty or sanctions by the company. It was appalling and scandalous that this would happen in our data science community. How could that have happened?
What followed the DC public statement was swift condemnation by numerous entities in the open source community: fastai, R-Ladies, SatRDays, RStudio, Data Carpentries and PSF (Python Software Foundation). As well, many members, particularly from the R community, publicly denounced DataCamp and encouraged learners to not take their courses which they had produced for the DataCamp platform.
The DC statement elucidated some details of the incident. There was clarification that it was not rape, both for the sake of the “target” as well as the “offender.” The prior lack of transparency of the incident had resulted in misinformation and unnecessary rumors. Because of the limited information, some in the public had misassigned the “offender” status to whichever male they associated with DataCamp prior to the actual name being released in the April 22 articles. As well, people had speculated on who the target was.
DataCamp’s Value and Liability
Like many tech organizations, DataCamp has brought value to the data science community as well as disappointment, agony, outrage, anger, regret, bitterness, frustration and confusion.
DC has an office in New York City and for those of us here, we see and interact with DC employees in person regularly at meetups, conferences and other events.
For people who hold incrementally differing thoughts on the spectrum of views, the DataCamp case has also created division and tension in the community which holds the values of education, inclusivity, community and open source near and dear to their hearts.
Part 2: Genuine Commitment
Start-ups Can Start With Diversity & Inclusivity for Future Success
There is much that DataCamp could have done differently. This section focuses on what DataCamp and other companies can do moving forward to transform their environment, that is to create a professional, diverse, inclusive and ethical environment.
People know when a company is “reacting” to a situation, “excusing behavior”, and working on damage control rather than “proactively” doing the gritty work of creating the environment that is needed. A true commitment means being transparent, courageous and making the difficult decisions over the advisement of biased decision-makers:
- executive team / management (exponentially more powerful than rank-and-file staff)
- investors (profit is priority)
- lawyers (what is legal is not always ethical)
- human resources (imperative to protect the company/management over the employee)
- public relations dire need to protect the public image (“looking good”)
It is commonplace to see companies conduct “independent investigations” and then disregard the findings and recommendations. Each of these parties has a sliced and nuanced approach which disregards ethics and long-term, unquantifiable consequences of toxic work environments.
Without the genuine commitment, implementing all of the recommendations following this paragraph *will not lead* to a diverse or inclusive work culture because the current organizational structures are not designed with any reliable checks and balances.
When the commitment to diversity and inclusion is genuine, the desired culture is created and only then can the harassment situations be prevented and lead to work environments where employees and businesses can thrive.
Freada Kapor Kline: Venture Capitalist + Sexual Harassment Researcher
Freada Kapor Kline is the Founder of Kapor Center / Leveling the Playing Field in Tech and founding team member of Project Include. Her degree is in social policy with a focus on employment policy. She wrote her Ph.D. dissertation on sexual harassment, and it contained 20,000 research subjects, equally divided between the two genders. She co-founded the first group on sexual harassment in the US in 1976 in Cambridge, Massachusetts.
Kline has worked in tech since 1984, beginning at Lotus Development Corporation. Kline is also a venture capitalist and she and her husband were seed investors for Uber. They both publicly shared their objections on how then-CEO Travis Kalanick was managing the toxic work environment.
Project Include’s mission is to to accelerate diversity and inclusion solutions in the tech industry. It was founded in May 2016 and its website provides a plethora of resources related to various dimensions that promote an inclusive culture for all companies.
Kline’s name should be as well known as those of tech CEO’s. She has a breadth of education and knowledge related to tech and inclusivity built over nearly five decades of experience.
If there is an incident, transparency is essential at all time points. The alternatives are rumors with false information. Often times, rumors are far worse than the reality of the situation and they can be damaging.
The names and photos of the people on the Advisory Board and Board of Directors and the executive team should be publicly available and easy to find.
DataCamp website’s About the Company does not provide any of that information.
Do not tolerate toxic leaders. Do not tolerate toxic employees.
Advisory Board Should be Diverse
The advisory board must be diverse with 50% representation by women. It is not possible to fairly and adequately represent women, people of color and other under-represented groups (URG) by a limited, few homogenous leaders. The URGs need to be at the table to receive requisite representation. The Advisory Board must be comprised with demographics representing the society to provide satisfactory advice.
The executive team and advisory board of DataCamp is not sufficiently diverse.
It matters who are around the founders table. It matters who the founders and VC are. Admit it.
A venture capitalist is not an “independent party.”
The discussion should be about who is missing from the team, not about what to do.
Venture capitalists can have a profound effect on diversity and inclusion by requiring their portfolio companies to commit to set and achieve diversity targets. VC’s should not be considered an independent party for sexual harassment or discrimination incidents.
Get a Chief Diversity Officer
Chief diversity officers, sometimes known as CDOs, are responsible for cultivating work or learning environments that encourage and support diversity. To be clear, it is not the job of the CDO to create an inclusive environment. It is every employee’s responsibility to contribute to that.
The CDO isn’t the one that is supposed to do all the work.
The Faces of the Company
Consider what photos and impressions are on the walls at work as well as on the company website and other marketing material.
DataCamp’s homepage shows 3 white men. It was founded by 3 white men, so it is an accurate reflection of what the founders believe represents the company. However, if the company wants to continue doing business in the heterogenous American society as well as other countries, their vision needs to include women and people of color.
Supporting, Training and Mentoring
Companies can cease justifying the lack of women at the higher ranks with the “pipeline problem” excuse. The current employees can be trained and mentored. Some company has to do it.
Women and URGs can be supported in these ways:
Listen to them, let them speak
- Advocate for them, highlight their work, amplify their voices
- Mentor and sponsor them
Recognize that business and mentoring happens outside of the office such as lunch or golf and that employees who are excluded from these social events are at a disadvantage at work.
These actions that support, train and mentor employees will retain them at the company as well.
Figure out how you will make the next 5 women and people of color in your organization CEOs and executives.
Hiring with an Intersectional Lens
Build a team that reflects your customer base: Latin, Asian, Black, disabled, age groups. Intersectional diversity include various dimensions of diversity: gender, race, disability, sexual orientation and more.
On hiring, rewrite job descriptions so they are more welcoming to people of URGs. Research has shown that various descriptions attract different groups of candidates.
Reach outside of conventional networks to other groups to reach a variety of candidates.
Publicly report diversity metrics and share achievements, challenges and setbacks.
Use an Intersectional Lens for diversity.
Tie Compensation and Bonus to an Inclusivity Dimension
Tie bonuses and raises to an inclusivity dimension based on employee annual review.
Intent doesn’t matter. Impact does.
Pay Equity Transparency
While exact pay does not need to be publicly shared, salary band widths are possible to share. Publicly state that pay equity is a priority and practice for the company, and then actually do it.
Full-time Human Resources
Start-ups prioritize technology, swag and retreats, but not Human Resources. HR is essential for all businesses, including start-ups. It is as necessary as a bathroom, the phone and internet access. For smaller companies, they are able to access HR experts via an outsourcing company such Xenium HR. Any venture capitalists and investors should require that, regardless of the size of the company.
DC was operating for almost 5 years without an HR representative.
To reiterate, It is possible to implement some or all of these initiatives and still have a toxic work culture if the genuine commitment is not there. True change begins with the mindset and then actions.
More on Project Include
Project Include has recommendations and resources on the following sub-topics:
- Defining Culture
- Implementing Culture
- Guide to Writing a Code of Conduct
- Employee Lifecycle
- Compensating Fairly
- Providing Feedback
- Training Managers
- Investing in Sponsorships
- Resolving Conflicts
- Measuring Progress
- Leading as VCs, educators, and employees
Part 3: Public Discourse
As demonstrated by the organizations that have boycotted DataCamp (fastai, R-Ladies, SatRDays, RStudio, Data Carpentries and PSF (Python Software Foundation)), the data science community values diversity and inclusivity.
What has been surprising is the conversation on social media platforms have received a strictly binary label as acceptable/supporter or not acceptable and which has shut down any dialogue for differing viewpoints.
R-Ladies community was the driving force in bringing this incident to light. After DC issued their first public statement on April 4, 2019, R-Ladies Global issued a statement on their disapproval of DC. That initial statement was issued on 08-Apr-2019 and then revised on 11-Apr-2019.
The initial statement is unavailable.
The statement was revised after this discussion on twitter:
I’m extremely disappointed with how R-Ladies has handled this situation. I agree with the intent of not tolerating sexual harassment and with promoting places that are positive environments for women. https://t.co/dy7TJOVEWz— Jacqueline "I have a book" Nolis (@skyetetra) April 9, 2019
Putting out statements on behalf of a group of 10,000 women that suggest they represent the span viewpoints of the women in the group is disingenuous, and it puts people in the group who don’t agree with the statement in an uncomfortable place.
I think the DataCamp product is a great resource for aspiring data scientists who can’t take transitional routes—especially women. Also, some of the best data scientists I know work at DataCamp, including multiple women and many allies.
I’m conflicted about how to best handle interacting with DataCamp, but I don’t agree with the R-Ladies stance that you should go as far as removing DataCamp stickers from your laptop.
This R-Ladies stance been especially difficult for me as a gay trans woman who was so excited to become a part of the R-Ladies community as part of my transition. Literally one of the first things I did post announcing was joining the slack. Seeing this statement hurts so much!
It makes me feel that because I don’t agree with the statement that was make by an unknown set of people representing the whole group, I am not welcome in it. That’s so hard for me because of how much I’ve felt I’ve done to join it and the friends I’ve made in it.
In closing, I want to emphasize that while I don’t condone how DataCamp has handled this or think sexual harassment should be tolerated, I wished R-Ladies, with such influence and responsibility, would take a more nuanced view and consider the for breath of women affected by it.
Addendum: to be specific, I wish that R-Ladies had:
1) been specific in who was making the statement and not make it on behalf of the whole organization.
If it was meant to represent all members, the group should be renamed because it’s not a group solely about being an R Lady who uses R and Also Agrees with a Few People who are at the Top
2) have a position that specifically addresses both the many underprivileged women who depend on the tool. Suggesting some other far less feature-full tools as good enough isn’t fair to those women
3) have a position that addresses the women working at DataCamp. These women have careers that rely on the work they are doing to succeed, and not explicitly weighing their needs seems like an unfair treatment of women.
R-Ladies reconsidered their position:
Thanks for your suggestion, @skyetetra’s remarks were part of our reasons for clarifying a few things (thanks @skyetetra!). However, we prefer having a readable best version at any moment hence our not marking changes.
The revised statement is available: R-Ladies Global’s disapproval of DataCamp
I have spoken to dozens of people in the data science community who disapprove of DC’s handling as well as the community pressure to respond with a specific viewpoint. One example, but not limited to, is the R-Ladies Global statement. Many in the general data science community also were tentative until further information was available.
People have privately confided that they have felt pressure to issue statements from affiliated organizations which they do not feel comfortable supporting.
But, their voices are largely absent in the public discourse.
It is possible for any/combo/all of the following to co-exist:
- Compassion for the woman who experienced the sexual harassment
- Disapproval of DataCamp’s handling of the incident
- To hold DataCamp responsible and not want to destroy the company
- Want and demand change at DataCamp (and other companies) for the better so an incident like this does not occur again
- Compassion for the employees of DataCamp, most of whom are not included in executive team discussions or influential
- Understanding that not all DataCamp employees were aware of full details of the incident
- Right of DataCamp employees to choose their place of employment
- To hold DataCamp responsible, disapprove of its handling of sexual harassment incidence and continue use of the product
- Referring to the incident as “sexual harassment” / “uninvited physical contact” / “inappropriate behavior” / “sexual misconduct” over “sexual assault” and still disapproving of it.
- …many other combinations and options
It is sad that this incident happened. It is commendable that a community collaborated to bring the incident into the public light.
Ironically, it is also sad that many people in the very communities created to increase inclusivity and underrepresentation do not feel comfortable expressing their opinion and have opted to not participate in the discussion out of fear of being verbally attacked. These unspoken opinions would be of tremendous value to achieving a resolution, preventing further incidents and ameliorating current organization situations that are unsatisfactory.
One community member shares:
That [bias] was the most frustrating thing for me and the whole twitter rampage… it just is not a space to have a reasonable conversation and we all lose if we cannot have a reasonable conversation.