Statistical modelling and data analytics
Many organisations require advanced data analysis but do not have the in-house technical expertise. Since we are highly skilled in both traditional statistical and more modern data science techniques, we are able offer our clients both methodology advice, in addition to conducting the analysis itself. We have extensive experience in survey data analysis, including the automation of such analysis, and are highly proficient in all the relevant software in this area, e.g., R, SPSS, SAS, SQL, Python Pandas.
Advanced statistical analysis of survey data
We conduct all forms of bivariate and multivariate analysis of survey data, including Segmentation, whether using Decision Trees or unsupervised Clustering algorithms, Regression and other Linear Modelling. If required, we can produce data tables, assist in report writing and interpretation of results, and provide tailored instruction in all areas of statistical analysis.
We have a keen interest in Key Driver Analysis, in particular the application of Kruskal’s Algorithm which we believe the be most suitable for dealing with the problem of multicollinearity.
See https://jumpdata.co.uk/software-development/key-driver-analysis-tool/
When appropriate, we develop and implement new and innovative methodologies when analysing such data, e.g., new choice models, different weighting approaches et al. As we believe in collaboration, we ensure that the client is intimately involved with our rationale at each stage.
From Our Work
Business intelligence, simulation and forecasting
We are experts in the development and analysis of effective KPIs for business and marketing activity. We use historic and current data sets to construct data-driven forecasts that enable organisations to make effective decisions, and derive which KPIs are the most effective to track over time.
We have built forecasting models in various software that continue to make highly accurate predictions many years after their original implementation. We have accurately predicted sectoral spend on a national level that have been widely reported in the media.
From our work
Automation and streamlining of data flows
We believe all tasks that can be automated, should be, and have a proven track record of the simplification of data flows. This is of particular importance with regards to tracking surveys and other repeated analyses.
We have automated mass production of data tables and various types of survey analysis, e.g., mimicking and extending stepwise regression. We have migrated many repeat surveys into normalised databases to allow simple comparisons across sectors and time periods.
