Despite general improvements in the amount of data available, official statistics are often not up-to-date. This makes the data, painstakingly collected at a huge cost to the exchequer, almost irrelevant to policy-makers
Over the past decades, the world has seen an increase in the availability of development statistics. The international development discourse has increasingly focused on results and goals, prominently illustrated by the Millennium Development Goals and corresponding national development goals. This focus on measuring development has improved statistical data availability.
But, despite improvements in the amount of data available, official statistics in developing countries is often not up-to-date, and reported with lower frequency. In particular, statistics on important welfare measures — poverty, health, and unemployment — are many years old.
For example, the World Bank, now has official, comparable poverty numbers for 98 developing countries between 2005 and 2010, as compared to less than 80 countries 10 years ago. But despite improvements in the quantity of data collected, the frequency and speed at which such data is available to policy makers have not improved. It often takes years before key indicators and micro data sets are available at international level.
Even while working at national level, recent household surveys data sets are more than three years old. The problem is not only the frequency of surveys, but also the lag between collection and data dissemination that takes more than a year. There are several reasons as to why data is not available to policy makers in time. Some are due to capacity, budget resources and even politics.
However a key reason for the relative slowness and low frequency of data collection is the method of collection and production: Pen and paper household surveys — sometimes referred to generally as pencil/pen and paper interviewing. There have been some improvements in the data processing stage, with the use of scanners for automatic data entry, but overall the process remains similar to what it has for many years, as enumerators go from door-to-door to collect data with paper questionnaires.
Moreover, statistical systems are mostly focused on large survey exercises focused on long, complicated questionnaires rather than monitoring key indicators. Two areas which have seen some improvements in the frequency of reporting is — prices (inflation figures) and some measures of food security.
A wide range of potential data collection methods and sources are now available as a result of the sweeping improvement in technical hardware (lower cost and better technical features) and connectivity (mobile ownership and network coverage) in the developing world. These approaches have been implemented to allow faster data collection, either as a substitute for smaller surveys formerly done by pen-and-paper interviews, or as a complement to achieve fast data on a selection of key indicators in near real-time.
For example, ‘computer assisted personal interviewing’ is an umbrella term for a set of survey technologies that substitute traditional paper surveys with electronic forms on tablets, smartphones, mobile phones or computers. Traditionally, the main advantage of CAPI has been that it skips the data entry stage when collecting surveys on paper. CAPI is not new, but has been considered expensive and impractical in many developing country contexts. Now, that’s changing.
Ensuring quality data also requires automated data processing. In this context, the cloud has enabled the computing model to shift from traditional data centers to publicly shared computing infrastructure; yet, applications leveraging this new computing model can experience performance and scalability issues, which arise from the hidden complexities of the cloud.
The most reliable path for better understanding these complexities is an empirical approach that relies on collecting data from a large number of performance studies. Armed with this performance data, we can understand what has happened, why it happened, and more importantly, predict what will happen in the future.
However, this approach presents many challenges, in the form of data management. We attempt to mitigate these by fully automating the performance measurement process. Concretely, we have developed an automated infrastructure, which reduces the complexity of large-scale performance measurement processes by generating all the necessary resources to conduct experiments, to collect and process data and to store and analyse data.
It is important to remember that timely data is only one of many elements necessary for responsive Governments, and that “knowing in time” does not necessarily mean “acting in time”. Even in situations with complete information and perfect, up-to-date statistics, there are often capacity, institutional and political limitations that prevent addressing the problems at hand.