Welcome to this project's final report! This report shares the outcomes, impact and learning from the grantee's project.

The knowledge of marginalized communities is the knowledge of the majority of the world. This project has explored an approach for centering the knowledge and expertise of marginalized communities in the Wikimedia movement in 2017. In partnership with Whose Knowledge?, two pilot communities (Dalits from India and the US, and queer feminists from Bosnia and Herzegovena) have led the way in mapping their own knowledge, creating Wikimedia content, and speaking powerfully about their truth and experiences. Together, we're exploring what it means to do this work in ways that have been at times both joyful and painful. The Wikimedia movement has much work to do to focus collective efforts on the knowledge and communities that have been left out by structures of power and privilege. Allyship is key.

“	There are also much longer and much more embedded battles to fight in spaces like Wikipedia. Like other indigenous knowledge, much of Dalit knowledge and history is in oral knowledge and primary sources. However, platforms like Wikipedia have very poor infrastructure for the representation of oral histories. Together, we have been working to "legitimize" Dalit oral history within white and Western pedagogical structures. Needless to say, these efforts have not been easy.	”
— Maari, Equality Labs

Goal 1: In partnership with community leaders and scholars, build a v1 map of existing gaps and opportunities around knowledge created by, and of, the two pilot communities: Dalits and women’s human rights defenders in the global South.

• Success! Knowledge maps exist for 2 communities.
  • Dalit map: in spreadsheet, in Wikipedia Requests
  • Okvir map: in Wikipedia-focused spreadsheet, and in a more broadly focused timeline

Goal 2: Clean up and systematise the raw mapping data into an easily searchable, editable database, with potential for language localisation and visualization of marginalized knowledges.

• Mixed success+failure, and some de-prioritization.
  • Through the mapping process, we learned what data is most useful to gather, what to leave out, and some basic ways to structure it that make sense to either Wikipedia, or to marginalized communities themselves (and we've uncovered some places where these two don't always easily intersect). This is a useful foundation.
  • But, the first solution we tried out for housing data (Wikipedia Requests) doesn't meet our communities' needs without further customization. That customization has been difficult to move forward in the absence of good code documentation.
  • Ultimately, this aspect didn't turn out to be a highly prioritized need from marginalized communities we were working with in this phase. Spreadsheets were ok for the first phase, and when this becomes a stronger need to help pilot communities achieve their goals we can revisit it again - long-term storage will certainly become an issue in future phases.

Goal 3: Support 2 edit-a-thons with the focus communities - in partnership with other individual Wikimedians and groups - to create or improve content on Wikimedia projects that are highlighted by the mapping process.

• Success! Content has been created and improved using this model.
  • At 5 Dalit history edit-a-thons in April-July 2017 (3 in India, 2 in the USA), 80+ community members and Wikipedian allies improved 185 articles, created 21 new articles (e.g. Grace Banu, and uploaded 27 files to Commons, all of which has had over 12M pageviews.
  • Okvir editing output on Wikipedia so far has been smaller-scale. After learning from both the successes and backlash from Dalit History Month, the Bosnian group has been compiling sources and laying the groundwork for a few strong articles (e.g. Lepa Mladjenovic) in multiple languages, while focusing larger-scale content production on Kvir Arhive, their oral history archive.

Overall aim: build and document a model that can be used, refined, and adapted for use by these and other focus communities working with Whose Knowledge? in the future, as well as by other Wikimedia groups aiming to address systemic bias in partnership with marginalized communities.

• We think of this report as a v1 model to be improved over time, and remixed into better forms for different contexts.
• Another version of this story/model is documented here.
• See the "future opportunities" section for what we think v2 looks like.

Planned measure of success (include numeric target, if applicable)	Actual result	Explanation
Data on knowledge gaps/opportunities produced by 2 focus communities is accessible online.	Dalit and queer feminist knowledge from Bosnia is available online under free license.	See Project Resources section.
Focus community testers are able to search, edit and add to data in their own language/country.	Spreadsheets + Wikipedia requests sort of allow this, but not with any elegance.	Deprioritized in first phase, to be revisited later.
User feedback from target communities and a focus group of Wikimedians working on systemic bias demonstrates usefulness of the approach and documented model, including the specific ways to map knowledge with rather than about marginalised communities.	Conversations with pilot communities, allied Wikipedians, and those who attended our panel at Wikimania 2017 indicates the approach is generally useful (can a measure of success be that Wikipedians cried in our panel?). Specific feedback from pilot communities has helped us learn what does or doesn't work in our initial documented resources. We haven't, however, gathered systematic feedback for public consumption to demonstrate this.	A more concrete survey beyond our current informal conversations might be something to consider in later phases, particularly after our next collective sprint on toolkits for other marginalized communities to use (see next steps section).
40 pieces of Wikimedia content created or improved (measurable increase from baseline in content about the Dalit community and women human rights defenders)	At least 185 articles were improved, 21 articles created, and 27 Commons files uploaded via Dalit History Month.	We beat our own expectations! That said, the queer Bosnian community has not prioritized edit-a-thons or other large content-creation pushes on Wikipedia at this time. And how much of the Dalit content will persist on Wikipedia long-term also remains unclear, because so much Dalit knowledge has been kept out of the sources and privileged perspectives that some of Wikipedia's most active editors prefer.

File:Dalit History Month Profile.jpg

This project began with 2 communities:

1. Dalit History Month with Equality Labs, a Dalit group based in India and the U.S. Equality Labs has been part of a collective organizing Dalit History Month for several years, with an aim to share the contributions to history from Dalits around the world.
2. Kvir Arhiv in Bosnia with Okvir - Feminist LGBTQI group in Bosnia and Herzegovina. Okvir has been collecting digital histories for a queer archive in Bosnia-Herzegovina, beginning with stories from activists and others about the 1990's war.

“	Our goal is to explore, document and make visible narratives and counter-narratives regarding security, gender and sexuality of LGBT*IQA persons and HRDs during the times of war and transitional justice period (from 1990-s to 2016).	”
— Azur, Okvir

• Each community brainstormed a list of important people, events, issues, etc relevant to their communities.(sample framework)
• These were then organized into spreadsheets, noting which topics were already covered on Wikipedia. (example of Dalit spreadsheet and LGBTQI Bosnia spreadsheet)
• Note: Some knowledge had to be redacted from the public versions of these maps, for security/safety reasons of both marginalized communities.
• Community organizers reviewed existing Wikipedia articles for quality and perspectives, noting where improvements were needed, and where content was missing.
• Community organizers reached out to community scholars to find reliable sources for use on Wikipedia. The sources that communities consider reliable often differs from the sources that Wikipedians consider reliable, so we spent a lot of time talking about sources all together!
• Community organizers used the maps to prioritize worklists for creating Wikipedia content. (examples of Dalit worklists)
• Note: Mapping takes significant time and energy by the organizers from each community! Additional funding for this was provided by another grant from another funder.

• Spreadsheets seemed to work ok for both communities to generate first maps.
• Afterwards, we cleaned up one spreadsheet for upload to Wikipedia Requests to test it for longer-term storage. Harej did the first bulk upload of Dalit knowledge, and built a bulk uploader for others to use in future
• Note: in present form we don't see Wikipedia Requests as a working solution for long-term storage of mapped data in a format that's super accessible and useful to marginalized communities. (See Project resources section below for more learning on this)

• Whose Knowledge? provided mini-workshop sessions with key organizers in each community.
• As needs emerged, we co-created resource pages and checklists to support the organizers.
• Community members gathered together at edit-a-thons to work on creating Wikipedia content.
• Wikipedian allies joined each event, supported content creation during and after events, gave feedback on sources both before and after articles were written, and advised when gatekeeping from a more experienced Wikipedian turned distinctly un-civil.

“	As a Dalit software engineer, being in a space that empowered us to create our own content was deeply inspiring. I felt I had ownership over a story so seldom told by others. I found histories and stories that were buried and forgotten. These are stories that get muted in the loud barkings of mansplaining. I learned about people like me. I learned about Grace Banu, another Dalit Software Engineer, who went through her computer science program without owning a computer. The editathon told me that I mattered, and that I had power over my own voice. I hope to continue to support more editathons in the future.	”
— Selvi2017, new editor

Our key organizing principles for pairing multiple experts and allies remain:

• Centering the expertise of marginalized communities themselves, with marginalized community organizers leading in key areas like knowledge mapping, prioritization/worklists, participant outreach, and edit-a-thon facilitation.
• Inviting community scholars who can help with sources.
• Inviting Wikipedian allies who have expertise in supporting, creating and curating online encyclopedic knowledge.

Mapping

• Framework for mapping your knowledge
• Queer feminist knowledge map from Bosnia - spreadsheet
• Timeline of queer history from Bosnia
• Dalit history knowledge map - spreadsheet
• Dalit missing knowledge map - Wikipedia Requests

Note - Here's some of what we learned would be needed to improve Wikipedia Requests as a long-term solution for storing mapped knowledge:

1. grouping/tagging/sorting lists for easier readability (right now we can find all DHM content in 1 unsorted list only),
2. ability to add priorities (useful for generating worklists for events coming from 1 bigger list),
3. documenting developer code well so that multiple developers can contribute (including newer Wikimedia devs).

Content creation

• Resource page to understand, practice, critique, and teach with Wikipedia
• Resource page on open licenses and digital repositories
• Edit-a-thon checklist
• Instructions for tracking edit-a-thon contributions
• Dalit History Month 2017 Meetup Page
• Dalit History Month Content Dashboard

Media, tech, stories, etc.

• Dalit History Month Visualizations - knowledge gaps/opportunities, process, and exploring quality from Dalit and Wikipedian perspectives
• Exploration of quality markers from Dalit perspectives
• Case study of Dalit History Month 2017 process
• Images on Commons: Whose Knowledge?, and Whose Knowledge at Wikimania 2017
• Azra Causevic speaking on challenges and opportunities of adding queer feminist knowledge from Bosnia online, on "Centering Knowledge from the Margins" panel at Wikimania 2017
• Thenmozhi Soundararajan speaking on challenges and opportunities of adding Dalit knowledge to Wikipedia, on "Centering Knowledge from the Margins" panel at Wikimania 2017
• Stan Rodriguez and Michael Connolly Miskwish speaking on challenges and opportunities of adding Kumeyaay knowledge to Wikipedia on "Centering Knowledge from the Margins" panel at Wikimania 2017
• Original audio snippet from 2017 "Centering Knowledge from the Margins" panel where Dalit and Native American speakers talk about their interactions with Wikipedia
• Wikipedia Requests bulk uploader

Note - Here's some instructions from the Wikipedia Requests developer about bulk uploading in the current system:

1. Spreadsheets you upload through this interface should have six columns: Page language code (e.g. put ‘en’ for English Wikipedia), Page title, Brief summary of the request, A longer explanation of the request, Relevant categories, Relevant WikiProjects
2. Example: I put each request in the “Dalit History Month” WikiProject, as well as sub-projects based on the topics given e.g. “Dalit History Month/Politics and Legislation.” If the page already exists, then the request will be sorted by those WikiProjects as well.

• Learning Pattern: Centering Marginalised Knowledge

Trying to go broad and fast, instead of focusing on deep and slow.

• Example: Wikipedia can be a really antagonistic and re-traumatizing space for marginalized communities, organizers, and activists who have had their knowledge questioned and erased in many other venues over many generations. Working on many articles at once created extra opportunities for some of the marginalized organizers to experience the kind of on-wiki harassment that's very familiar to many Wikipedians, and unfortunately is also all-too-familiar to many women, LGBTQI people, people of color, etc who live on the broader internet. This year, we've deepened our commitment to encouraging marginalized communities (including ourselves!) working on Wikipedia to 1) start with a small set of content to focus on learning with and shoring up, thereby limiting the sheer number of potentially re-traumatizing encounters, and 2) make room for some additional safe, joyful spaces for sharing knowledge outside of Wikipedia when things get too painful. These are mitigating steps until the Wikimedia movement has better solutions for grappling with the larger issue.
• Example: Last year, we thought we would begin with large-scale open source knowledge “maps”, that would be led by global surveys of marginalized communities. We quickly realized this was entirely the wrong approach. Until we built trust with a few communities and learnt what their frames of reference for knowledge and “expertise” were, we couldn’t hope to create a global survey that would be useful in different contexts. We may come back to a global survey - but not until we have a far better grasp of what multiple specific knowledge maps for different communities looks like, and what they imply for others like them.

Working with multiple pilot communities has given us a lot of ways of doing this work, with an approach that centers communities and their needs. We are planning a convening with our core community organizers in 2018, using the Book sprints methodology, to bring together our shared experiences and expertise and create set of tools and resources that other marginalized communities can re-mix and re-use, based on their contexts. This will be one next test to see how our approach can be scaled thoughtfully.

Expense	Approved amount	Actual funds spent	Difference
Design and tech	$18,000	$15,000	$3,000 not needed in this phase
Community organizer	$6,000	$6,000	0
Project coordination	$12,000	$12,000	0
Hospitality (food, printing, etc)	$500	$996.52 + $78.51 in-kind WMF merchandise	($575.03) more was needed than planned, for extra events
Travel	$8,000	$7890.61	$109.31 left over from travel
Fiscal sponsor overhead	$3,340 + $269	$3,609	0
Total	$48,109 USD	$45,574.64	$2534.36 underspend

Do you have any unspent funds from the grant?

• Yes, $2534.36 remains unspent - we invested less than expected in tech for this exploratory phase.

If you have unspent funds, they must be returned to WMF. Please see the instructions for returning unspent funds and indicate here if this is still in progress, or if this is already completed:

• We'll be sending back funds, in coordination with WMF grants admin staff.

Did you send documentation of all expenses paid with grant funds to grantsadminwikimedia.org, according to the guidelines here?

• Yes

Did you comply with the requirements specified by WMF in the grant agreement?

• Yes

Is your project completed?

• Yes

“	We learn something new each day, and are grateful to be able to do this work with love, joy and amazing partners! We hope Wikimedia funders will continue supporting explorations in new directions like this, keeping in mind that different sorts of metrics, timelines and rhythms are needed for different contexts. Let's keep expanding the boundaries of what it means to be Wikimedians, as we make space for the majority of the world to share their knowledge. Thank you for your support!	”
— Siko and Anasuya, Whose Knowledge?

“	Some of what we are doing and encountering has often felt so profound when placed in historical context. Dalits and other caste-oppressed people were not allowed access to reading, writing, or learning for millennia....It is deeply momentous to be able to bring our people, and bring their voices and find a pathway to self-representation on platforms like Wikipedia. Working on the internet in this way is not just self-representation but really, acts of protest and transformation!	”
— Maari, Equality Labs