# TICTeC@Taipei - 9/12 Morning - 201C

:::info
Please join us to take notes by clicking on the pen icon on the left upper hand.
:::

## Civic Tech for Following the Money

> * Impacts of Budeshi.ng on Nigerian Public Procurement, **Ifeoma Judith Onyebuchi**, Public and Private Development Center (Nigeria)
> * The Impacts of Open Contracting Reforms in Ukraine and Slovakia,**Gavin Hayman**, Open Contracting Partnership
> * Serenata de Amor: Tracking Congress People's Expense Claims in Brazil, **Ana Schwendler and Jessica Temporal**, Data Science Brigade (Brazil) Serenata de Amor: Tracking Congress People's Expense Claims in Brazil, **Ana Schwendler and Jessica Temporal**, Data Science Brigade (Brazil)
> * Moderator: Julia Kloiber, Open Knowledge Foundation Germany

### The Impacts of Open Contracting Reforms in Ukraine and Slovakia,**Gavin Hayman**, Open Contracting Partnership

Why public procurment? - it's where the money and the power is.
It's govs #1 corruption risk.

Open contracting helps to identify and fix problems.
It's not e-procurement, not a top-down mechanism for contract renegotiation

OCP works with governents. It's an open data schama describing how to release docs and data at each stage of the contracting process as marching readable, flexible, extensible user-friendly data.

Responds to user needs (value, integrity service delivery...)

Uses structured JSON data. One contract can contain many procurements, states of the process.

Results from public contracting experiment in Slovakia. Possible, highly-feasable, highly beneficial. Direct corruption exposed. Increase of competition (bill doubled) public interest went up.
Similarly encouraging results from a e-procurement experiment in Ukraine (see [Prozorro](http://prozorro.gov.ua)).

***

### Serenata de Amor: Tracking Congress People's Expense Claims in Brazil, **Ana Schwendler and Jessica Temporal**, Data Science Brigade (Brazil)

Brazil ranked 79 in the transparency index => Can do some efforts on corruption. Brazil is already good on transparency (9th).

How come there's so much corruption and openess at the same time? Daily expenses of representatives. The expense records are public but not easily accessible.

SdA - started a 3-month crowdfunding campain. Rosie (AI bot) processes expense records and uses machine learning to identify suspicious records that stand out from the expected trends.

Used hypothesis and an agile approach to be make the campaign work is a restricted time.

**Result:** a dashboard where we can find the expenses, displayed in a nice readable format. A community is now alive (~600 members on Telegram.)

The project has a twitter bot account that tweets suspicious transactions when it finds them. This has raised awareness of the project and promoted a dialogue between parliamentarians and people on Twitter.

See serenatadeamor.org/en for more.

### Question (both sessions)

**Q (to David)**: How do you update the data?
People can amend the contracting recods. You can track who made the last change, when etc...

**Q (to SdA)**: How sensitive, specific is Rosie. How's the classifier doing?

We build the classifiers ourselves. It's unsupervised ML. We cross-refference data. We focus on solving the problem, not the algorithm. We come with hypothesis and solve it with code.

**Q (to David)**: Do you approach the governement directly, to you talk to civic society?

Not always government-led. Tends to be that way, often born in civil society and is picked up by the governement once they see the value (e.g. case study on Nigeria from Ifeoma). Sometimes driven by Government, which tends to be slower, but too early to tell which one is most successful long term. Efficacy depends on whether or not the process involves public participation and isn't just transpency for transparency's sake.

***

### Impacts of Budeshi.ng on Nigerian Public Procurement, **Ifeoma Judith Onyebuchi**, Public and Private Development Center (Nigeria)

Public and Private developmet Centre.
Procurement monitoring. Why tracking public procurment? It's the key, enabler of what we can do as a nation.
Even though the law requires to provide PP data within 7 days it often takes over 2-3months.
80% of curruption in Nigeria initiate from public procuments cases.
It's impossible to track each project from inception to end.

People who pay taxes should be able to track how the money is spent.
Need of a tool that will track information on a per-project basis.

The tool can compare budgeted amount vs contracted amounts for each projects. It also geolocates the locations of the projects and show their advancement.
Enables people to go onsite and check if that corresponds to the reality.

Outcomes: improved public resource management, greater sector reforms, less fraud & corruption, more accountability.

Current steps: scale this system to the states, train journalists, integrate mobile-based plateforms instead of solely internet based (people have phones).

## Technology for Civic Participation/Engagement
> * Findings of the Taiwan Open Government Report, **Po-yu Tseng and Mg Lee**, Open Culture Foundation (Taiwan)
> * Keyboard Participation: An Analysis of ‘g0v’ for Collaborative Participation by the Open Source Community, **Ting-Yu Cheng** National Taiwan University
> * The Impacts of PopVote in Hong Kong, **Virgile Deville** Democracy Earth (France)
> * Digital Skills Training for Better Citizen Participation in France, **Orianne Ledroit and Pierre-Louis Rolle**, The French Digital Agency
> * Moderator: Luke Bacon, Open Australia Foundation (Australia)

### Findings of the Taiwan Open Government Report, **Po-yu Tseng and Mg Lee**, Open Culture Foundation (Taiwan)

[Here are the slides](https://docs.google.com/presentation/d/19mSuLJz1uWJf8i4bRFHMVe6viVmFa1IDb0WY1EximBA/edit#slide=id.g23ce012c0a_0_0)

See more about the report and download the PDF [here](http://opengovreport.ocf.tw/en/#tab-0).

Taiwan Open Gov Report - 4 chapters:
- Law and Policy
- Open data
- civic participation
- civic tech collaboration with government

**i-Voting system** Requires thousands of signatures to introduce a canditate.

**Case Study Shazidao**

The [*join* campain](join.g0v.tw) - people can propose, talk through and supervise proposals.

**vTaiwan** - discuss regulations associated with the Internet. Allows discussions between gov and civic society on relevant subject matters.

Tools used: pol.is, Sayit, hackpad, youtube.

### Comparison of the different plateforms

####Where
online/offline. Not always easy for some people to participate online.

####Who
3 categories
1) builds the tool
2) sets the agenda
3) can discuss/give opinions

For example, for vTaiwan gov+community do ==1/== and ==2/== general public do ==3/==

#### When

Follow this model:
Problem analysis -> Planning -> Legistlation -> Implementation -> Evaluation

#### Impacts and insgihts

* Online and offline meetings has to mix
* Impact depends on case more than on tools (e.g. Uber case on vTaiwan)
* Political support of high leaders is key
* Public attention may bring more participation but can hinder quality

#### Statement from Authors of Blulu Metrics

Blulu is a Creative Commons licensed work and anyone can use it without our permission. We did NOT participate in this report. Blulu was originally designed to vulgarize the practice of public deliberation, NOT intended to be used in a formal report as the methodology.

報告的「參與」章節所大量使用的 Blulu 模型，原作者聲明：

呂家華的聲明 https://www.facebook.com/lu...
藍一婷的聲明 https://www.facebook.com/ET...

摘要如下：
1. Blulu 採用創用 CC BY-NC 授權釋出，任何人都可以拿去改來用
2. 創用 CC 是一種開放授權，任何形式的改作都不需經過原作者審查，作者也不替改作品背書
3. Blulu 的兩位作者完全沒參與 OCF 「開放政府觀察報告」的架構設計，僅由藍一婷簡單回答改作過程中遇到的問題，呂家華則完全與報告無關
4. Blulu 模型原本是設計來傳遞科普知識，不是設計來套用在正式報告，套用在正式報告的改作，需由改作者自行評估

什麼是 Blulu？請參考： https://blulu.tw/

***

### The Impacts of PopVote in Hong Kong, **Virgile Deville** Democracy Earth (France)

CivicDataHK - tech professional UX designers, front/backend developpers. Specialize in polling, chat-bots, strategic voting prediction and civic referendum.

San-francisco based fundation working on liquid democration. Now worknig on a software to facilitate these techniques.

Referendum gathered 1/4 of the voting population in HK to use PopVote. DDoS hacking attempt amusingly helped gain attention.

Can vote on telegram. People need to verify their identity using phone number, ID card etc... 4 attacks where detected, targetting organizers email accounts, the polling station wifi, to byass phone certification and the DDoS attack.

Rumours (a smear campaign) of user data leaked really harmed the reputation of PopVote, even though they were unfounded.

The depency on Telegram is not ideal because of the software and depency on the repsonsiveness of their team.

The social smart contract - paper they're currently working on.

slides: frama.link/popvote

***

### Keyboard Participation: An Analysis of ‘g0v’ for Collaborative Participation by the Open Source Community, **Ting-Yu Cheng** National Taiwan University

Joined g0v during the Sunflower movement in 2014. Personal challenge of understanding how g0v works (too many geek words!!).

#### How does g0v thrive?
1) Plausible promise : ppl can initiate their initiatives
2) Effictive tools: e.g. Slack Hackpad, Github. Based on interactivity, decentralisation. This helps raise participation
3) Acceptable bargain: guaranties there will not be monopol. Incourages to trap volunteers.

#### The model Sharing, cooperation & collective action
from *Here Comes Everybody Clay Shirky*: Sharing, cooperation & collective action

**Sharing examples** - convert online paper into digital information, visualise information
**Cooperation examples** - Participants reach consensus through project contribution
**Collective action examples** - Initiate project as a policy to accomplish the change. But was blocked by the government.
Neutral conversation: government communicate with citizens through projects or .platefrom

Citizens can join at any time of the process.
Engagemet can lead to objection, neutral conversation or acception by the gov.

#### Limitations

It's not public voting yet. How to bridge the digital devide?

***

### Digital Skills Training for Better Citizen Participation in France, **Orianne Ledroit and Pierre-Louis Rolle**, The French Digital Agency

Purpose Work on equality and economic growth. Provide people with digital opportunities.
Civic tech tools are being used in the 3 areas of :
1) Digital inclusion
2) citizens empowerment
3) smart territories

40% of french citizens are afraid of making online administrative procedure.
There is a perception that it is not a powerful/trustable way to make impact.

Can digital inclusion be the antidote?

We work on local and nationwide initiatives. There is a need of a lot of groundwork ; a one-size fits all solution for providers, users and trainers.

#### The product

APTIC « Plateforme de Gestion » - will be open-source. One can see their progress with digital skills, know the locations of the nearest training centers, get certifications etc...
Starting with 4 cities, 4 taxcenters, 6 training centers.
Cities where on very different areas (La Réunion, ... )

#### Outcomes

* Civil sertvant and digital trainers learned how to address both fiscal and digital issues.
* turnout was ower than expected
* significant gender gap (60% were women)
* 55% people expressed the will to learn new skills

Twitter:
@pierrlouisr
@orianneledroit
source: github.com/societenumerique

### Questions

**Q: Do your know initiatives designed for countries where the government is very closed (eg. Vietnam)**

Alice & Virgil - In HK the gov is working to track down any civic acktivity. Always think of the next step, something the government didn't expect. ANd work on increased participation.

MG Lee - Civic tech becomes famous only after some kind of crises (eg. g0v after the Sunflower Mvt).

**Q: What if the governement doesn't want to play the game? Was it a challenge for you to get their support?**

Po Yu - After the Sunflower movement the awareness of the people put pressure on the government. They HAVE to respond. Some are pretending to be open, some are really willing to help.
But even though we have vTaiwan, i-Voting etc... This is not enough. There is no genuine co-creation yet.