Analysis of spatial intelligence from community mobility reports and implications for policy options in Kenya and Africa

The new month of May, new trends in global COVID-19 numbers

The month of May has ushered in a new COVID-19 consciousness, notably in Africa. Having recorded surging numbers lately, Tanzania and Kenya epitomise this disenchanting reality in East Africa. These new shooting numbers have defied the previous comforting trends constrained within the limits of previous routine tests. By May 3, the cumulative global COVID-19 cases had risen to 3.5 million with an improved recovery rate of 32% and a case fatality rate of 7%. This trend makes it clear that the earlier projection here (link) of 3.7 million cases by the end of May is going to be exceeded by a wide margin. It also means that the previous model assumption of continuing to witness persistently low numbers in Africa as a latecomer in the staggering COVID-19 statistics has been invalidated.

The new global growth trend in cumulative COVID-19 cases has effectively entered a linear phase with a steep gradient, simulated here using

since April 18th, 2020. This new equation projects the cumulative global numbers to hit 4 million by April 9, 4.6 million by mid-May, and 5.9 million by the end of May. This considerable difference from the previous projections proves just how the numbers can change quickly over a short time for such a highly infectious pandemic.

Is Africa’s map of COVID-19 cases changing with increased testing rates?

The cases in Africa are rising with mass testing. The picture in Africa by May 3 featured lower case fatality rates than the global rate of 7%. Uganda and Rwanda had not recorded any fatality while Ghana (0.8%) still had a much lower case fatality rate than Kenya (5%). By the same date, the global recovery rate had increased to 31%, higher than Ghana’s (11%), but below the recovery rate in Kenya (35%), Rwanda (44%), and Uganda (61%).

Uganda’s increased COVID-19 testing rate doesn’t take away from Ghana’s achievement

Kenya started on mass testing from the beginning of May and consequently registered a significant increase in a population-and-time-normalised testing index from 7.2 to 8.1 tests/million/day. Testing continues to exert its importance as a source of strategic intelligence to combat COVID-19. The COVID-19 datasets from Africa until May 3 have shown that the countries fronting more aggressive testing protocols for their populations have also displayed better crisis management. By this date, Ghana had posted 72 tests/million/day, far ahead of Uganda (17.6), Rwanda (17.3), and Kenya (8.1). The testing index for Korea (118) was much higher than these Africa’s records, whereas countries like Spain, Italy, Germany, the UK, and the USA have lately been registering much higher than 250 on this measure. Targeted mass testing is proving critical to averting the complacency of ignorance in the fight against COVID-19.

Can Africa calibrate containment policy action using citizen science?

The leading question here is:

Given the glaring resource limitations most countries are facing, how can Africa utilise location-based intelligence crowdsourced from the public to combat COVID-19?

Amidst the surging COVID-19 cases within community clusters, riding the wave of enhanced mass testing such as the one Kenya kicked off from May 1, leaders should spare a thought for the crucial role of spatial intelligence. This is the location-based intelligence provided through modern satellite-based navigation and mapping technologies, including mobile mapping. Community mobility reports acquired using satellite-based technologies are a strategic asset in combating COVID-19. Such reports are publicly available, owing to the initiative by Google to provide aggregated and anonymised sets of data from mobile users with location features.

At a crossroads between tightening containment measures and easing up on selected economic sectors, country leaders will be great beneficiaries of the science-policy interface that public-facing technologies can offer. For a rich mix, a careful balance of expertise drawn from science, technology, and humanities is key. It is worth noting that several countries have started putting in place taskforces on easing lockdowns towards a post-corona era. Intelligence from space, complete with remote sensing capabilities, may just be the critical ingredient needed to complete the armoury for a winning combat as the pandemic’s wave resurges with relaxation in social distancing measures. Policy and privacy questions still prevail, but it remains a case-by-case challenge for each country to make a prudent decision for public good.

To seasoned researchers, these mobility reports constitute the embroidery of progressive citizen science, the authentic fabric of participatory, economical, actionable and community-facing decisions. In the hands of result-oriented leaders exercising public goodwill, such actionable intelligence springs to life through the full implementation of the ensuing decisions, causing impactful societal progress.

Previous COVID-19 articles in this thought-leadership series focused on the metrics and evidence-based models meant to influence policy, planning, and behaviour change. This article shares examples on how the location-based intelligence (spatial intelligence) from community mobility reports can inform containment policy and strategy for combating COVID-19. Such knowledge aids in calibrating region-and-sector-specific interventions. Data on sixteen (16) countries across different continents have been used for illustration purposes.

Spatial datasets were drawn from Google LLC “Google COVID-19 Community Mobility Reports”. The baseline is the median value, for the corresponding day of the week, during the period January 3–February 6, 2020.The baseline has been used to compare the community mobility reports from March 15 to April 26, 2020. Such datasets crowdsourced from the public inform citizen science, made even more prolific with precise location codes obtainable from today’s widespread GPS receivers. The reports tracked community mobility trends across geographical areas over time. The place categories considered were retail and recreation, groceries and pharmacies, parks, transit stations, workplaces, and residential areas. These categories are key to social distancing efforts and access to essential services.

Basics of satellite navigation systems for people in a hurry

Recent advances in global navigation satellite systems (GNSS) have greatly improved the positioning accuracy available for civil applications, public health being a key area. These space satellites include the USA’s GPS satellites, the most cited among other constellations such as Russia’s GLONASS, Europe’s Galileo, and the Chinese BeiDou.

GNSS satellites use highly accurate atomic clocks, the speed of light, and principles of trigonometry to provide reliable data on geographical locations (represented by coordinates) as well as speed measurements. These useful data are sensed by receivers held by various users on the Earth’s surface, as long as they occupy favourable locations under the open sky to receive adequate signals from four or more space satellites. Modern smartphones are equipped with GNSS receivers, qualifying them as today’s leading examples of disruptive technological convergence.

Global trends showing substantial reductions in mobility for retail, recreation, and transit

The table below summarises the 16 country cases. Against the baseline data, high reductions in community mobility around retail and recreation including restaurants over the study period have been realised in Spain (92%), India (86%), France (83%), and the UK (78%). The decrease in mobility trends around the transit stations for these countries has also been high, from 64% in the UK to 82% in Spain. Germany and Finland have shown an interesting anomaly in the study group. For the two, there has instead been an increase in some non-residential mobility categories: visits to parks including beaches and gardens (78% and 56% increase, respectively) and a 3% increase in visits to grocery and pharmacy in Germany.

Africa’s high mobility reductions in retail and recreation including restaurants have been registered in South Africa (72%), Uganda (65%), and Rwanda (63%). The three countries have also registered high mobility reductions around transit stations: South Africa (78%), Rwanda (75%), and Uganda (73%). The satellite-based technology has confirmed relatively small reductions in figures for Tanzania (33% in retail and recreation; 27% for transit stations) and a relatively slight increase in the residential area category of 9%. This figure falls way below the corresponding residential mobility increase in Egypt (15%), Kenya (20%), South Africa (22%), Uganda (23%), and Rwanda (24%). The variance in containment policies among these countries could largely explain their differences against the baseline values, with Tanzania as an outstanding case of policy divergence.

Kenya and the case of a digital divide across counties

Only twelve (12) out of the 47 counties in Kenya show some fair to good level of technological readiness in terms of sufficient mobile connectivity to generate remotely sensed data towards boosting location-based intelligence in the fight against the spread of COVID-19. This observation is based on the community mobility data generated by Google from civilian satellite-based (GPS) services between January 3 and April 26, 2020. The datasets further provide insights into the level of mobile digital access to expect across the vast country with a wide urban-rural inequality in technology infrastructure, among other indicators.

Nationally, Kenya has realised a decrease of 50% in community mobility trends around retail and recreation including restaurants, 47% around transit stations, 39% around grocery and pharmacy including farmers markets, 33% around parks including beaches and gardens, and 19% to workplaces. Residential areas have registered a boost with an increase of 20% in community mobility trends.

This trend implies that before any major intervention, the uptake of mobile solutions including the prospects for delivering online educational services may favour only a quarter of the 47 counties (see the table below). As sensed remotely from satellites, only five counties generated enough data over the three-month study period to support reporting on all the six community mobility categories: Nairobi, Mombasa, Nakuru, Kajiado, and Kiambu. Kisumu, Uasin Gishu, and Machakos generated enough data for five categories, leaving out residential areas. Kilifi, Kitui, and Meru provided enough mobile data in less than five categories.

The community mobility reports show that between March 15 and April 26 and with respect to the baseline data of January 3–February 6, 2020, all the twelve Kenyan Counties in the table realised a major beating in the area of retail and recreation including restaurants, registering a fall in mobility ranging from 41% in Machakos and Nakuru to 57–58% in Mombasa and Nairobi.

Nairobi, Mombasa, and Kisumu have also recorded the most profound change in mobility trends. These three counties have particularly registered a major decrease with respect to the baseline data in the following areas: retail and recreation including restaurants, visits to parks, gardens and beaches, and transit stations.

Mombasa and Nairobi have realised a significant increase in mobility trends around residential areas (+20 to +22%), so far only topped by Kajiado (+24%) and Kiambu (+24%). Changes in mobility around transit stations have been substantial in Uasin Gishu (-59%), Kitui (-57%), and Kisumu (-56%), with Machakos registering a change of -44%. Kilifi, Nyeri, and Uasin Gishu have also registered a higher reduction in mobility in the category of grocery and pharmacy including farmers markets.

Conclusion

Globally, COVID-19 containment measures have been reflected in a major reduction in community mobility in the categories of retail, recreation, and transit. These trends have been observed against the baseline satellite data before the strict containment measures took effect. Cases of strict containment measures have led to higher reductions, in Spain and India for example. Residential areas have received an increase in community mobility, mostly by about 20%. Workplace mobility data have generally recorded more reductions in countries with higher levels of development and a fair proportion of formal sector jobs. Their higher levels of automation and digitisation could explain this observation.

In Kenya, the COVID-19 containment measures so far imposed have been effective in limiting community movements. Retail and recreation places, transit stations, and grocery and pharmacy including farmers markets have realised the most significant reduction in mobility trends — mostly by more than 40%. Residential areas, as expected, have gained in mobility trends, by about +20%. The smallest reduction, mostly barely 20%, has been in work-related mobility trends; this is a further testament to the necessity of travelling to work for the majority to make a living in Kenya’s largely informal economy.

The intelligence aggregated from space constitutes the citizen science which is instructive to the governance of COVID-19. In terms of social distancing, community compliance is key, either through informed choice or strict policy enforcement. Allowing some room for work-related travels, as has been the case with the curfew, remains a wise choice since remote working is far withdrawn from the majority in the country as per the ground reality revealed here.

For the education sector, which holds the fate of a critical mass of young citizens, the spatial intelligence from this analysis exposes the inequality in digital access, a digital divide which favours only a quarter of the counties and more so their urban headquarters. Until a major shift in digital access happens, hopes for a successful nationwide delivery of online learning to students on forced holiday across Kenya must await a delayed fruition date. The satellite-based community mobility reports show evidence of being resourceful in informing location-specific COVID-19 containment policies. Kenya and similar countries in Africa, therefore, need to weigh options carefully to reap the dual benefits of security in curbing surveillance capitalism and utilising the resources of modern spatial technologies contain COVID-19.

Data Reference

Google LLC “Google COVID-19 Community Mobility Reports.”
https://www.google.com/covid19/mobility/ Accessed: 03 May 2020.

Worldometer (online). https://www.worldometers.info/coronavirus/#countries. Accessed: 03 May 2020.