All
Since there are too many overlapping curves we break the charts up by country and then by hosts within country. The spreadsheet can be found here, each country has its own tab. The spreadsheet of the raw data is here.
Burundi
PingER monitors two hosts in Burundi. They are a government host www.assemblee.bi and the University of Burundi (www.ub.edu.bi). Neither appears to have marked changes around March-April. Also there is little information on interventions etc. in https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Burundi. According to https://www.imf.org/en/Topics/imf-and-covid19/Policy-Responses-to-COVID-19 "Measures taken to minimize the risk of the pandemic breaking out in Burundi have been very limited: The population has been instructed to follow some basic rules of limited social distancing and frequent handwashing. Hand sanitizers and water for handwashing have been installed in public places."
Eritrea
PingER monitors one host in Eritrea, it is www.eritel.com.er. A telecom and Internet provider. Unfortunately, there was no monitoring at what was probably the critical period.
Ethiopia
There appears to be a sustained increase in IPDV in Ethiopia around Mar 16 when schools were closed and public gatherings were suspended. Looking in more detail at the 3 hosts monitored in Ethiopia there are 2 government hosts (www.amharainfo.gov.et and www.most.gov.et) and a university. For www.most.gov.et PingER data is missing for this time period. Below we see www.amharainfo.gov.et and www.stmaryuniversitycollege.edu.et had a sudden rise in IPDV between Mar 12 and Mar 15. See below and the Excel spreadsheet.
According to https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Ethiopia#March_2020 the first Covid-9 case in Ethiopia was on Mar 13. "On 16 March 2020, the office of the Prime Minister announced that schools, sporting events, and public gatherings shall be suspended for 15 days. On 8 April 2020, the Council of Ministers declared a five-month long state of emergency in response to the growing number of coronavirus cases. The state of emergency was approved on 10 April by the parliament".
Kenya
There are 3 hosts being monitored by PingER in Kenya. Two are government hosts both in the capital Nairobi, and one (www.daystar.ac.ke) is the Daystar University in Kampala. Looking at the chart of IPDV below it appears that IPDV dropped by several 10s of a percent sometime between Mar 21 and Mar 28.
From https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Kenya we see: On 15 March 2020, President Uhuru Kenyatta directed that the following measures to curb COVID-19 be implemented:
- "All schools and higher learning institutions be closed by Friday March 20, 2020.
- "Government and businesses people start working from home; except essential services.
- "No congressional meetings – weddings, malls, night clubs, churches, limitation of visits to hospitals.
- The Health Minister Mutahi Kagwe banned all social gatherings including religious gatherings on the same date.
- All bars were to remain closed from 22 March, with restaurants allowed to remain open for takeaway services only.
On Mar 25 a nationwide curfew was announced to start on Mar 27. The curfew was extended by 21 days on Apr 25.
Looking at the charts below it is seen that there is a noticeable reduction in the two-week average IPDV that coincides with the interventions noted.
Kenya details by host
Looking at the averages, standard deviations, (standard deviation/averages) plus the derivatives of the averages, standard and (standard deviation/Average) for 30 days before and after each date, we get the charts below. It is seen that within about a week of the interventions the derivatives, in particular, took a sharp rise for all 3 hosts.
Madagascar
PingER monitiors 2 hosts in Madagascar. There are www.cnaps.mg a private national insurance fund, and www.mefb.gov.mg the Ministry of the Economy and finances. Both hosts monitored in Madagascar are in the capital city of Antananarivo which has a population of 1.6 million or about 6.1% of the population of Madagascar. According to https://www.mining.com/sumitomo-halts-mines-in-bolivia-madagascar/ there was a curfew starting Thurs Mar 20 and 2 cities went into strict lockdown on the same date. Within the two weeks before and after Mar 20 there does appear to be a very slight reduction in the variability of the PingER IPDV measurements.
Malawi
PingER monitors three hosts in Malawi. They are the Parliament of Malawi in the capital Lilongwe, (www.parliament.gov.mw), the Kamuza College of Nursing at the Queen Elizabeth Central Hospital, in Blantyre (www.kcn.unima.mw) and the University of Malawi - Malawi Polytechnic (www.poly.ac.mw) in Blantyre.
From https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Malawi the main interventions that might be expected to impact network utilization were:
- Both public and private education institutions were closed from 23 March
- President Peter Mutharika confirmed the country's first three cases of coronavirus disease 2019 on 2 April
- On 4 April, suspended all public meetings, gatherings and conferences;
- From 8 April, the Lilongwe City Council placed a temporary ban on the opening of bottle stores, pubs and night clubs, sports clubs and gyms, street processions, wedding receptions, bridal showers, birthday parties and any other social gatherings or public events.
- On 9 April, Government announced more restrictions on entertainment, public events, markets and religious gatherings; • Schools, colleges and universities remain closed since 23 March;
- On 14 April, President Mutharika announced a 21-day lockdown starting Saturday 18 April at midnight.[17] However, on 17 April, the Malawi High Court temporarily barred the government from implementing the 21-day lockdown following a petition by the Human Rights Defenders Coalition
There is a sustained increase for a few days in IPDV seen by the University of Malawi on Mar 23, however, a similar rise was seen the week before. Similar rises were also seen by www.parliamen.gov.mw. There is a reduction in the overall magnitude and also the variability of IPDV for all hosts for 2 weeks after Apr 19 following the proposed lockdown.
Looking at the 14 day rolling average and standard deviations of the IPDV measurements for www.parliament.gov.mw chart below, we see a roughly factor of 2 drop in the average and standard deviations around the proposed April 18 lockdown.
Mauritius
PingER makes measurements to two hosts in Mauritius. One is www.utm.ac.mu at the University of Technology Mauritius in the capital Port Louis. The other is www.groupeunion.mu the Groupe Union Mauritius, in Riche en Eau, Mauritius. They track one another closely.
According to https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Mauritius on Mar 19 a sanitary shutdown was ordered starting Mar 20, then, on Mar 22 a curfew order was ordered effective Mar 23. Following the University shutdown there appears to be a drop in the IPDV. Then following the curfew there appears to a persistent increase for about a week and a half. Whether or how these changes in the IPDV may be related to the interventions measured is unclear.
Mozambique
PingER monitors four hosts in Mozambique. They are www.bancomoc.mz the Bank of Mozambique, www.rm.co.mz Radio Mozambique, www.issm.gov.mz the Insurance Supervisory Institute of Mozambique (ISSM), and www.uem.mz the Universidade Eduardo Mondlane. All are in Maputo the capital.
The first case was Mar 22, on Apr 8 there were 7 new cases for a total of 17 of which 9 were local transmissions. There is no mention of interventions in https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Mozambique. There appears to have been an ~50% reduction in IPDVin the 1st week of April for www.bancomoc.mz, and a large drop in IPDV for www.uem.mz in the 1st week in March.
Rwanda
PingER monitors three hosts in Rwanda, they are: www.auca.ac.rw the Adventist University of Central Africa in Kigali, www.lmis.gov.rw the Rwanda Development Board in Kigali and www.rwandaparliament.gov.rw the Rwanda Parliament also in Kigali. They all track one another fairly closely.
According to https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Rwanda the main intervention that would impact Internet traffic would appear to be on 21 March, officials announced a two-week lockdown. Both public and private employees are to work from home, under strict measures. There is a small dip in IPDV around this time, however, it is not evident that this is due to the lockdown. Also, there is a larger dip two weeks later. Also on 9 April, three new cases of coronavirus were identified, out of 720 samples examined in the previous 24 hours. This brought the total number of confirmed cases to 113 people (of which seven had recovered)
Looking at the 14 day rolling average and standard deviations for www.rwandaparliament.gov.rw we see the chart below:
Sudan
PingER monitors five hosts in Sudan. They are: www.abs.sd the Agricultural Bank of Sudan in the capital Khartoum, www.bashaer.sd Bashaer for Energy Services in Khartoum, www.cert.sd Sudan CERT (Computer Emergency Response Team) in Khartoum, www.iua.edu.sd the International University of Africa in Khartoum, and www.sustech.edu the Sudan University of Science and Technology in Khartoum.
According to https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Sudan, the first case was Mar 12. There is no indication of any interventions. According to https://reliefweb.int/report/sudan/sudan-covid-19-situation-update-15-may-2020 a three-weeks long lockdown imposed in Khartoum since mid-April was on 9 May extended for nearly two weeks as part of containment measures to prevent the spread of coronavirus. www.iua.edu.sd stopped responding on Apr 1.
www.cert.sd experienced a dramatic sustained increase in IPDV in the week following the first case.
Tanzania
PingER Monitors 4 hosts in Tanzania. Three are universities: the Open University of Tanzania (www.out.ac.tz) in Dar es Salaam, the former capital, that retains most government offices and is the country's largest city, principal port, and leading commercial center; the Sokoine University of Agriculture (www.suanet.ac.tz) in Morogoro a town with a population of 315,866 (2012 census) in the eastern part of Tanzania, 122 mi) west of Dar es Salaam; the State University of Zanzibar (www.suza.ac.tz) in Zanzibar town. The other host is related to education: the Tanzania Commission for Science and Technology (www.costech.or.tz) in Dar es Salaam. Of these only www.suza.ac.tz appears to notice any substantial change in responsiveness. In particular, there are big dips at weekends. Also, the peaks before Mar 21 are larger than those after.
Looking at https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Tanzania we find:
The first case reported was on Mar 16. On 17 March, Prime Minister Kassim Majaliwa announced a range of measures, including closing schools. This corresponds strongly with the drop to near zero for www.suza.ax.tz starting on Mar 18. The first death was recorded Mar 31st , in Dar es Salam. On 21st May, the President announced that colleges will reopen and form six secondary school students will return to school from 1st June, sports will resume from 1st June.
 |  |
 |
|
The large spikes in IPDV for www.costech.or.tz between Feb 18 and Feb 26 are not understood. They are also seen if one looks at the hourly minimum RTT, average RTT, maximum RTT and losses.
Looking at other PingER IPDV changes and averaging the IPDV over a 14 day window we get the chart below for www.costech.or.tz. There is a noticeable reduction in the average IPDV and its standard deviation following Apr 4. However, we can find no interventions around that time.
Uganda
PingER monitors two hosts in Uganda. One is the Victoria University in the capital Kampala (www.vu.ac.ug), the other is a commercial host mail2.starcom.co.ug also in Kampala.
From https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Uganda
On 18 March, public gatherings including places of worship, pubs, weddings, music shows, rallies and cultural meetings were suspended. Mar 22 1st case confirmed. On 25 March, public transport was suspended for 14 days. Only private cars with not more than three occupants are allowed on the road. On 26 March, police and other security personnel were heavily deployed in all city suburbs, slums and along the streets to enforce the president's directives. A few people who breached the government's restrictions were arrested. On 30 March, the President declared a nationwide curfew from 7 pm to 6:30 am, which would run for 14 days to prevent the spread of the disease.
These interventions appear to correlate strongly with the rise in IPDV at the Victoria University and to a lesser extent for mail2.starcom.ug.
Zambia
PingER monitors five hosts in Zambia: www.deczambia.gov.zm the Drug Enforcement Commission in the capital Lusaka, www.cooperativecollege.ac.zm the Zambia Cooperative College in Lusaka, www.talk90point5.co.zm the Millenium Radio in Lusaka, www.parliament.gov.zm the NationalAssembly of Zambia in Lusaka, and cobra.unza.zm the University of Zambia in Lusaka.
According to https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Zambia, the first 2 cases were on Mar 18, the first death was Apr 2. On March 17, 2020, the Zambian government announced that all schools, colleges and universities will be closed on Friday 20 March in the wake of the coronavirus pandemic. As seen below there does appear to be an impact (reduction in IPDV by about a factor of two for about a month) on the PingER IPDV for www.cooperativecollege.com.za and www.talk90points5.co.zm
Taking rolling 14-day averages of the IPDV we get the charts below. There is a strong correlation for the Cooperative College Univesity as it closed on Mar 20. The correlation is less strong and starts the week before for the Millenium Radio www.talk90points.co.zm:
Zimbabwe
PingER monitors two hosts in Zimbabwe. Hellenic school (https://www.hellenic.ac.zw/) and a commercial site www.paynet.co.zw
From https://en.wikipedia.org/wiki/COVID-19_pandemic_in_Zimbabwe we note the first case was Mar 21 and the first death was Mar 23. There were the following interventions:
- On March 23, President Mnangagwa announced Closure of bars, nightclubs, movie houses, swimming pools, and sporting activities.
- On March 27 the government announced that the country would go into a nationwide lock down for 21 days from the 30 March onwards.
The chart below shows that towards the end of March the IPDV of both hosts drops and smoothes out. This appears to correlate with the interventions mentioned above. To be more specific, following the spike on Monday Mar 30 the 2-week average for (excluding the spike) drops to between 1.06 and 1.3 (including the spike the average is about 1.5-1.6) for www.paynet.ac.zw.
