Apollo is an open-source data management system that is designed to support citizen election observation and other large-scale structured data collection using field-based observers. NDI hosts Apollo for partners through its DemCloud cloud-hosting service.
Please contact us for: use cases, approximate cost and a timeline for implementation
Citizen watchdogs play a critical role in validating political processes, but to be convincing they must back claims with data. Elections are one of the foundations of a legitimate democracy when the official results truly represent the will of the voters. Systematic election observation requires large amounts of structured information from hundreds or thousands of observers and determining what it means – fast. Developed by Nigeria-based TimbaObjects in conjunction with NDI’s Elections team, Apollo aids the management of observers, verification of collected information, and automated aggregation for analysis. Apollo has been implemented in dozens of countries, including Azerbaijan, Cote d'Ivoire, Kenya, Liberia, Malawi, Mali, Moldova, Morocco, Nigeria, Tanzania, Tunisia, Uganda, Zambia, and Zimbabwe.
Overview of Features
- Used in dozens of countries for systematic observations of critical elections.
- Useful for management of any large group of people collecting structured form data, such as monitoring government services, medical centers, or collecting survey data from citizens door-to-door.
- Automated processing and aggregation of data from easy-to-use forms.
- Accepts submissions via SMS, smartphone app, or web entry.
- Provides real-time information on citizen-submitted reports.
- Built-in graphic interface makes it easy to build data forms.
- Ability to upload detailed observer information including contact phone numbers and location data.
- Permits analysts to verify data from multiple observers in the same location.
- Automated SMS response for incomplete or inaccurate submissions.
- Blast text messaging and observer management for consistency and validity.
- Proven ability to process thousands of data points in minutes.
- Ability to export information for further analysis.
- Multilingual interface compatible with non-Roman alphabets including non-Unicode fonts and characters.
Technical Specifications
- Language: Python
- Database: PostgreSQL
- Web server: NGINX
- License: GPL
- Distribution: GitHub
Apollo in Serbia - Door-to-Door Survey
Center E8 had a mission: to find out what were the most important problems facing their fellow Serbians and demonstrate democracy in action by helping leaders solve them. This required knowing what the people of Serbia really care about, and E8 had an ambitious plan to find out by knocking on two thousand doors and interviewing people. These face-to-face conversations are some of the most effective ways of understanding people’s needs, and can be far more representative than telephone surveys. However, organizing hundreds of volunteer supporters of Center E8 in this effort, spread out across the entire country, would require sophisticated data collection.
Working with NDI, the youth community center’s leadership set up a central database and shared a link to the Apollo system with their volunteers. Since the system is designed as a “Progressive Web Application,” Apollo can be saved on a smartphone as if it was an app, making it easy for staff to access. The web application was particularly useful in rural areas where cellular networks are less available because data can be recorded offline and then uploaded once connected to the internet. For first-time volunteers learning any new software, particularly for data collection, can be challenging. “We all started by saying ‘I hate Apollo’, laughed Kristina Petronijević, the Center program manager for this initiative. But with training, practice, and some bug fixes, it worked great. One of the advantages of open source software, like Apollo, is that software developers are able to easily make changes needed to fix the bugs that came from the wide range of phones being used. Before long, Apollo and the Center team started smashing through their goals. Their first target of 2,000 citizens was quickly reached, and they raised the bar to 10,000. Eventually, E8 contacted 27,000 citizens. The group is now planning to double that number again, gathering valuable information on the highest political priorities of over 50,000 Serbians. This is data that no newspaper, university, or political party in the country has, all aggregated in the Apollo data management system.
Apollo had a range of features that helped Center E8 leadership in their mission, enabling them to see whether they were hitting their interview targets, while coordinators were able to communicate with the data collectors as soon as problems emerged. The geolocation of reports mapped the locations of their interviews in real time, which proved a real inspiration for the team - volunteers were empowered by seeing that their individual dots fed into such a large collective sea of achievement. Kristina commented that the E8 team was grateful to have saved enormous amounts of money by building on Apollo rather than buying commercial products or developing their own solution. The team paid a few thousand dollars for software improvements and bug fixes and about $100/month for hosting, which was relatively little for such a large operation.
The Center E8 team is now working with municipalities to share the real issues their residents face, helping leadership know what the most important concerns may be. Bringing data from Apollo to the conversion, as the team was able to do, changes the ways policymakers think about a civil society advocacy group such as Center E8, and will lead to a new focus on the most important problems that Serbians face today.
Apollo in Côte d'Ivoire - Sample-Based Election Observation
POECI conducted the country’s first-ever sample-based citizen observation for the 2015 Presidential election, which used Apollo to collect data on the conduct of the election to provide credible, independent analysis on the quality of the process. POECI observers used cell phones, SMS messages, and standardized checklists to provide real time data on key aspects of voting and ballot counting, including the opening and closing times for polling stations, presence of appropriate electoral materials, successful deployment of new biometric authentication technology, and adherence to the electoral code. Their observation also enabled an independent estimate of the election results.
On election day, POECI deployed observers to a representative sample of 755 polling stations across the country, covering each of the 14 districts, 31 regions, and 107 departments of Côte d’Ivoire. Observer reports, in the form of coded text messages, were aggregated in Apollo, for rapid review and processing by trained statisticians. POECI ultimately received 5,285 observer reports via 38,000 text messages, which included 396,375 data points – and 100 percent of the data expected from every single observer.
POECI observers reported that the process of voting went smoothly; nearly all polling stations had the right supplies and staff and followed the correct procedures. The observation statement noted scattered incidents, including intimidation or violence towards election officers, but no nationwide trends raised questions on the overall credibility of the elections. The evaluation of the process as free, fair, and credible, along with the verification of the outcome, helped candidates and political party members accept the results and advise their supporters to remain peaceful, win or lose.
In the 2016 parliamentary elections and 2018 regional and municipal elections, POECI continued to improve its capacity to deploy observers, collect data and communicate seamlessly across the organization. They have used Apollo and integrated communication tools to mass message observers for activity reminders, as well as to customize survey question responses from groups of observers on the fly. POECI is planning to continue to use Apollo as they look ahead to monitoring the presidential elections in 2020.