Time-based trends in (Quadratic Funding) Gitcoin Grants
A deep dive into the macro-informational database of the Gitcoin-supported grants from January 2018 to December 2021.
This submission contains only a high-level overview of our analysis and results. For you fanatics of statistical programming, the code script can be found here. We used elements of Natural Language Processing, exploratory data analysis, as well as statistical tests to uncover patterns in the data and discuss their implications.
The objective of the data bounty was to discover trends and gain novel insights into the Gitcoin (Quadratic Funding) Grants ecosystem using a dataset containing a lifetime history of grant applications and results.
To achieve this objective, we divided it into sub-objectives by coming up with the following questions:
- Which keywords are the most prominent in Grants titles and have they stayed prominent over rounds?
- To what extent does community engagement in projects vary per Grant category?
- To what extent does the amount of crowdfunded contributions (USD) differ per Grant category?
- How does the average number of contributors in each Grant category compare over rounds?
- Hypothesis: There is a significant difference between community engagement for big and small projects.
In this section, we will mainly talk about results within the plot that stand out or that directly answer our questions. Due to the magnitude of the outliers in the dataset, we zoom in on some plots to accurately show the trends and answer our questions. For a deeper dive and more plots, check out the code script on Github.
Gitcoin Quadratic Funding Grants are built and governed on the Ethereum blockchain and its community. So, it is not surprising that throughout the lifetime of grants, ‘ethereum’ is a prominent keyword in Gitcoin Grant titles according to our corpus of top keywords.
Looking closely at the frequency counts in Fig. 1, it seems that ‘panvala ’and ‘league’ go together as both have ca. 200 counts. The Panvala League is a sustainable community-owned fund built on, surprise, Ethereum. We can say also that many Gitcoin Grants were projects focused on developing ‘defi’ (ca. 275) on ‘ethereum’ (ca. 320) and also that a noticeable portion of the grantees were members of the ‘panvala league’.
Along with the crypto adoption, Gitcoin Grants has witnessed a massive growth in grantee numbers. From obtaining less than 150 grants per round, Gitcoin Grants experienced significant growth in Round 6 and again after Round 8. In Fig. 2, the less prominent sector Health enjoys its only increase in Grants in Round 5 (March 2020 to April 2020) from zero to nearly 40 grants. This spike is explained due to the extra funding allocated by Gitcoin to Public Health-based Grants during that Round towards covid-19 research. Figure 2 also shows that during the beginning of the pandemic, Grants focused on Community-, Decentralized Apps (dApp)-, and Infrastructure Tech development along have doubled and tripled in numbers.
The combined results from Fig. 1 and Fig. 2, tell us that most of the Gitcoin Grants since Round 5 have been for further growth and integration of the Ethereum community (Community), enhancing Technological Infrastructure (smart contracts) within the ecosystem, and dApp development. The love for dApps and community building continues to triumph when we look at the average community engagement in these categories. Fig. 3 shows the average community engagement over time for Grants by the number of unique contributors.
It’s clear from Fig. 3 that during Gitcoin Grant Rounds the preceding Round 8 Hackathon, the average community focused on ‘building Gitcoin’ grew from less than 200 unique contributors in Round 6 to 4000 unique contributors. Keep in mind this is a near 2000% growth in just 6 months (June 2020 to December 2020. Fig. 3 looks closer at the data and allows us to see how Infra Tech and decentralized Governance (dGov) achieved increased average community engagement since Round 8. If this is directly attributable to the Gitcoin Grants Round 8 (GR8) Hackathon seems highly possible given Fig. 1–3.
Prior to round 8, the Infra Tech category has received relatively less unique contributors compared to dGov or community-driven Grants. The average amount of unique contributors it has been obtaining from Round 8 onwards, makes it one of the most supported categories (Fig. 3). Figure 4 showcases how Infra Tech and dApp Tech-focused Grants went from totaling under 200k in crowdfunding until Round 7 to more than double from Round 8 onwards. The same plot depicts how the total crowdfunding for community-driven grants increase drastically due to the momentum of Round 8 and again during Round 12. During that period, NFTs have seen an increase in popularity which also explains the increase in NFT category Grants.
Round 8 (December 2020) of Gitcoin Grants, witnessed a significant increase in the amount of crowdfunding for Community, Infra Tech, and dApp — focused projects. External sources and sanity checks show this spike to be highly attributable to the Gitcoin Grants Round 8 (GR8) Hackathon after which the Gitcoin Grants community nearly doubled in numbers as seen in Fig. 4 & 5.
Another interesting aspect is that of the matching amounts. While the total amount in crowdfunding per category skyrocket due to the growth of the Gitcoin community, the matching amounts do not behave similarly. For example, a total of 400k in total crowdfunded USD for round 8 for Infra Tech dev (Fig. 4) does not coincide with a mere 100k matching amount (Fig. 5). Our correlation analysis in the next section also shows the matching amounts in relation to the crowdfunded amount to be varying per round.
So far we’ve explored each category for trends in keywords used in titles, total crowdfunding contributions, and community engagement over time. We saw that Round 8 doubled the Gitcoin Grants network and that the Quadratic Funding mechanism has effectively been able to sustain this growth ever since (Fig. 4 & 5).
For our last question, we investigated if community engagement statistically differs for big and small projects. In other words, do successful projects really have more invested contributors? This time, we use ‘project sentiment’ as a proxy for the community’s enthusiasm for a project and set the cut-off at 1k in total project funding as successful or not. We define project sentiment as the number of contributions relative to the number of unique contributors.
We act on the sensible premise that contributors donate several times to a project if they really believe in its potential. Because the data is not normally distributed, we can’t rely on classical statistic methods such as the famous T-test and confidence intervals. Therefore, we decided to take a nonparametric approach and use a Permutation test to estimate the difference in project sentiment and see if this difference is significant.
We hypothesize that there is difference between the average project sentiment for projects with more than 1k and project with less than 1k in total USD contributions. The result from the Permutation test shows that with p < 0.001, we should reject our hypothesis stating that there is a difference in average project sentiment.
Overall, Gitcoin Grants Round 8 and 12 seem to have had the most momentum (Fig 4 & 5). Since the beginning of that same year, the corona pandemic has been hitting economies hard so this explains the growth in Public Health grants during Round 5. Round 7 witnessed the biggest matching amount pool until that point, which logically speaking would lead to more Grants. Although, Figure 2 shows a decline in the number of Grants.
Also, the results from the permutation test and correlation analysis tell us that projects who have a higher contribution/contributor ratio are likely to result in more crowdfunding. Secondly is that during ‘special’ Gitcoin Rounds (such as 8 and 12), this effect is greatly strengthened.
To Gitcoin, we advise the obvious: keep organizing Hackathons, big events and allocating extra funding to emerging Grant categories to boost development in that area. Also, consider more collaboration with the Panvala League seeing that its members are prominent users of the Gitcoin Quadratic Funding model (Fig. 1).
From a total of 5906 entries, there were 716 entries with ‘undefined ’or ‘0’ regions. We propose fixing this issue for more accurate future (geo)- analyses. As crypto adoption, blockchain technology, and the Gitcoin community grow, there could be more categories and regions so our advice is to include more options and limit the freedom of not filling in data to ensure data accuracy. For example, when creating a Gitcoin Grant of the ‘Region’ box should be required and could include more options. We also propose including ‘project tags’ and ‘project stage’ and ‘project description’ in the dataset for in-depth trend analysis with modern methods.
This dataset only consists of ca. 6000 Gitcoin (Quadratic Funding) Grants, focused on building non-excludable and non-rivalrous digital public goods. Meaning the insights drawn here are not reflective of the performance of the entire Ethereum ecosystem and should not be used for non-educational purposes.