Ans Gerber

I'll be presenting on how we use ML for financial applications

11

1 Comment

In today's data driven world it is essential for everyone to be data literate. If you are looking to learn the fondational skills, check out the free courses by SAS and South African Statistical Association (SASA)

In a previous post, we showed Statistics SA’s estimates suggest that Cape Town is the only major metropolitan area to have recorded real house price increases since 2010. In that post we deflated nominal house price indices using headline consumer prices (CPI). In today’s post, we instead use owners’ occupied rent (an estimate of the cost of housing services for homeowners as if they were renting their own homes) based on the expectation that it would be more directly linked to housing market dynamics. CPI covers a wider range of goods and services prices, which may not necessarily align with housing market trends. One might expect, on the other hand, that owners’ occupied rent might provide a more accurate representation of the connection between property prices and rental markets over time, offering a clearer view of whether property prices are increasing disproportionately compared to rents. This insight is crucial for evaluating potential housing bubbles or issues related to affordability. Property prices have kept rising in Cape Town since the COVID-19 pandemic relative to owners’ occupied rent, whereas real property prices moderated when deflated by CPI. At a provincial level, Northern Cape real property prices rose faster than those in the Western Cape between 2010 and mid-2021, but have been declining since. See all the charts here: https://lnkd.in/eUKuXYKF Compiled by Almaro De Villiers

28

1 Comment

In today’s blog post, I repost my Business Day article in which I discuss what democratising data within and between organisations requires and what we need to do in South Africa to modernise the data ecosystem. https://lnkd.in/dbWMYtYM

8

1 Comment

Today my office released the annual Mid-Year Population Estimates, the first since the 2022 population Census. This has allowed us to review many of the inputs and assumptions that go into this important report which I hope, will continue to provide useful insights into the demography of South Africa. A few highlights of the report are that the population of South Africa has passed the 63 million mark as of the 1st of July, 2024. Approximately two-thirds of the population resides in Gauteng, Western Cape and KZN provinces. About 60% of the population is under the age of 35 but the 60+ age group continues to grow faster than any other broad age group at about 2.5% per annum. Our fertility rate is estimated at 2.41 children per woman in her reproductive ages (15-49) provided that current patterns prevail until women reach the end of this age group. This continues to confirm an ongoing downward pattern over time. A life expectancy at birth of 66.5 years means that we have recovered from the life expectancy decline due to COVID-19. All provinces except for Free State, Limpopo and Eastern Cape show a positive net migration between 2021 and 2026. #MYPE #demography #populationestimates Statistics South Africa Chantal Munthree

170

8 Comments

Daan Steenkamp will be presenting a summary of the comparative strengths and weaknesses of South Africa's data ecosystem and what should be done to address South Africa's widening data gaps - 15 April (16:00 SAST) Dr Daan Steenkamp is a co-founder and the CEO at Codera Analytics, an economic consultancy providing modelling, forecasting, and reporting solutions. Codera Analytics contributes to the democratisation of data in South Africa through its open data platform EconData. Join DAMA Southern Africa and Daan Steenkamp on 15 April at 16:00 SAST to learn more! RSVP link in the comments below. #damasa #bigdata #datascience #coderaanalytics

6

2 Comments

We've written a blog post about Mortality Statistics in South Africa, follow the link below or in our stories to give it a read: https://lnkd.in/dcJ3tyup - - #datascience #datasouthafrica #southafrica #deathrate #stats #ourworldindata #symbyte

1

Today's post shows three simple proxies of household's marginal propensity to consume in South Africa. While the ratio of household disposable income to consumption has been flat since 2008, the ratio of cheque deposits to time deposits and the ratio of compensation to retail sales have both declined. During this period, the household saving rate has been negative on average, so this decline would be consistent with higher consumer stress, weak consumer demand and borrowing capacity. Social grant income would be reflected in disposable income, which might be one reason why the blue line is flatter than the cyan line. https://lnkd.in/dqhSyDVA Riaan Grobler

12

Most of South African provinces' revenue come from allocations from the national government in the form of Provincial Equitable Share (PES) transfers and conditional grant allocations. The real per-capita value of these transfers have been under pressure and are expected to remain under a lot of pressure. EconData's South African Public Finance Module makes it possible to download time series of Budget Review concepts as time series and vintages from individual Budget Review documents. You can automate models and dashboards that use national and provincial fiscal data, whether you work in excel, R or Python, without having to manually copy data from individual budget documents. https://lnkd.in/dnT_ve38

20

Artificial Intelligence (AI), a transformative technology, is a key driver of global digital transformation. It offers many new opportunities to break down barriers to human development, enhance social inclusion, and align with the Sustainable Development Goals (SDGs). At the open online project launch on the "Development of South Africa's AI Maturity Framework (MF)", panellists discussed the need for a MF and the potential outcomes of the assessment, such as: informing holistic policy interventions, guiding investment decisions by both government and private sectors, establishing a baseline for AI maturity, mitigating intersectional inequalities, facilitating robust data governance, and providing recommendations for nurturing an ethical human-centred globally competitive AI ecosystem that balances innovation with safeguarding SA citizens from potential harms. Panellists Ms. Deshni Govender – GIZ South Africa, Lesotho & eSwatini Ms. Debra Setsiba – Department of Communication and Digital Technologies Ms. Shamira Ahmed – Data Economy Policy Hub (DEPHUB) Prof. Jacques Joubert – University of the Western Cape (UWC) Moderator Dr. Humphrey Brydon – UWC Council for Scientific and Industrial Research (CSIR)

3

1 Comment

On the importance of developing data assets and intellectual property for growth and welfare drawing on Codera's research: https://lnkd.in/dNEWuXU9

8

The South African Reserve Bank (SARB) publishes composite business cycle indicators to provide an early signal of what economic data imply for economic conditions. Analysts have been interpreting a pick-up in SARB's leading indicator as a sign that economic growth will recover this year. But do these indicators help to predict GDP? I find that the leading indicator is not a particularly good predictor of GDP. This means economists should be careful about using the leading indicator to make unconditional statements about the outlook for GDP.

17

6 Comments

Today's post provides evidence of a high degree of dispersion across price categories that comprise the consumer price basket in South Africa. We show that, since 2008, a large proportion of inflation outcomes for the detailed components that comprise the consumer price basket have been outside the Reserve Bank's 3 to 6 percent inflation target band. Both the mean and median outcome across very detailed specific components comprising the inflation basket have been above 5% since 2008. There has been high dispersion for several categories that are highly dependent on tradable products such as petrol (such as transportation), administered prices such as electricity, and some food sub-categories. High inflation dispersion has implications for the optimal inflation target and the optimal approach to inflation targeting. Read the full post for more: https://lnkd.in/dZ9nCXk8 Oliver Guest

18

2 Comments

Today's post by Aidan Horn presents two different aggregations of provincial infrastructure spending in South Africa. The first includes infrastructure planning, construction, maintenance and management and the second includes various administrative support such as associated corporate services, personnel, safety and various development and research and special programmes. Not accounting for inflation, there has been a large decline in infrastructure spending over the last 5 years (especially under our expanded grouping) and it is forecasted to remain flat over the forecast period. See all the charts in the post: https://lnkd.in/ee5VPs4j As far as we know, EconData is the only source of Provincial budget data for South Africa in an easy to use form for use in automating workflows using excel, R or Python. This is a goldmine dataset for research and policy analysis. Get in touch for a demo.

26

Part of the new CCT data strategy launch. Improving the way we build data products that add value in the City of Cape Town

93

4 Comments

Just released! Our latest #EconomicOutlook report offers a forward-looking perspective of South Africa’s post-election economy and business environment. It highlights insights from our leaders across key industries reflecting on their short-term outlooks. The report also looks ahead to 2025 and how South Africa could achieve more than 3% economic growth under a reform-focused scenario. Read the full report for more insights. #EconomicGrowth

8

The male-female population pyramid in South Africa historically showed a broad base, indicating a young population with high birth rates, but over time, it has become more column-like as birth rates have declined and life expectancy has improved. This shift reflects the impact of improved healthcare and social changes, leading to an aging population and a more balanced distribution between males and females in younger age groups. At young ages, there are more males than females, while at older ages, one can see that there are more females than males in the population. https://lnkd.in/d7JJE6ZJ Compiled by Johan Hanekom

90

6 Comments

CLV (or customer lifetime value) models are one of the most underrated applications of statistics and data science in South Africa. Stochastic models exist in many companies' assumptions (retention rate, transaction frequency, etc.) but are often “gut based” vs backed by science. These models are seldom backtested for accuracy and often exist in someone’s head or excel sheet, as opposed to in the hands of all decision makers. It is also one of No company is too small for CLV models, but some companies can take it to the next level with Machine Learning CLV, which requires more data. If you’re unsure how to apply AI at your company, CLV could be the best place to start.  - https://symbyte.tech/

2

My first article for The Conversation Africa explores the BER's long (70 years!) business survey history.

28

1 Comment