Technobunnies offers data consulting services. By looking after your data, we provide statistical analyses, data visualisation and data cleaning services. As demonstrated by our past client list, we are able to assist commercial clients, NPOs, private individuals or academics. As a registered vendor for UCT, postgraduate students, researchers and departments at the University of Cape Town are able to apply an existing grant to pay for the service.
Services
If you need someone to understand your data and its potential to provide information then get in touch and ask for a consultation. We can assist in the following areas:
- Research design
- Sampling design
- Questionnaire design
- Data collection
- Guidance for data capturing and verification
- Data cleaning and manipulation
- Statistical analysis of data
- Visualization and graphing of data
Reality Bites
Did you know that many of our scientists do not conduct the appropriate statistics? They attempt and fail quietly in offices with no network to reach out to, or they hand the job over to a statistician, who may network, but still does not conduct the right test or report the correct value. These 2 issues add up to at least half of the published scientific articles applying statistical methods containing statistical errors.
Common Mistakes
The most common mistakes reported in the research paper cited, are failures to adjust or account for multiple comparisons or even reporting a conclusion based on observation without conducting a statistical test (!), or use of statistical tests that assume a normal distribution on data that follow a skewed distribution. The most common reporting errors are unlabeled or inappropriate error bars or measures of variability, and failure to describe the statistical tests performed. Technobunnies is here to assist you in avoiding these common pitfalls. Even if the study cited above provides an overestimate that 50% of articles are riddled with error, then it is not a stretch (either statistically or mentally) to imagine that at least 1/4 of all studies suffer from such mistakes.
Battle plans
Technobunnies follow this methodology for most projects.
Just as no battle plan survives contact with the enemy, no data survives its intended journey to analyses. There is always something that wasn’t thought of, don’t feel bad if it happened to you, it happens at every level. The good news is it does not matter that you expected parametric data and didn’t get it, there are different statistical tests you can apply.
Do you want your data handled in MS Office? Your data is in rows and you need columns, that is a cinch to correct. You want to code or rank your data? A little dusting of VB, or some simple row-context IF statements can assist.
Comfortable using JS, Python, or R for your data handling? We provide the files so that you can read your code. Perhaps you have a huge data-set with a control and experimental group and you are sure you have something great, but wouldn’t know a t-test if it jumped and shouted boo?
Whether you are at the planning stage or ready to conduct your data analyses we can assist you in conducting the correct statistical tests, or simply consult and offer advice.
Pingback:University of Cape Town - technobunnies.com
Pingback:Data - technobunnies.com