class: center, middle, inverse, title-slide .title[ # Client Reporting: the Biggest Mistakes to Avoid ] .author[ ### Dr Thiyanga S. Talagala ] .author[ ### ] --- ## Client Reporting: the Biggest Mistakes 1. Making reports too long and monotonous 2. Not choosing right visuals 3. No summary of work 4. Reporting too frequently or infrequently 5. Spending too much time on report 6. Fully automated without thinking of the client’s needs --- ## How to Create a Consulting Report? 1. Make a title page 2. Include a table of contents 3. Include an executive summary: This is to ensure that your readers have a rough idea of the report’s content and encourage them to read on 4. Write an introduction 5. Methodology/ Data Analysis 6. Summary/ Conclusions 7. Appendix --- ## Tips for Writing a Client Report 1. Maximize readability 2. Write clearly 3. Be accurate 4. Write for your audience --- ## In-class: Errors on Your Reports --- ## 1. Client information Name, department, e-mail and phone no. of the client. --- ## 2. Problem to be addressed - Scientific background (field, subfield, topic) - Main question to be addressed from client’s perspective (write as one sentence or short paragraph) - Related previous studies by same investigator or other investigators in the same field - If applicable, obtain copies of papers, especially with examples of previous data analyses, and always ask the client which statistical methods were used before for the analysis of similar problems. --- ## 3. Study Design - No. of independent cases or units - Experimental or observational study? - Sampling procedure – iid? repeated measurements? clustered sample? dependencies? Is sample representative for population? randomization? - Special designs (ANOVA, Split plot, Case-Control, Sequential, Equivalence study etc.) - Planned or unplanned design - Check with client for (often hidden) biases or dependencies in the data or measurements --- ## 4. Measurements and Variables - Which/how many measurements were made per subject/unit? - Types of variables, physical units and statistical distributions - Variation of measurements, measurement process, possible transformations - Predictor and response variables --- ## 5. Preliminary Analysis - What kind of analysis has the client done, if any? - Exploratory data analysis, basic plots, data cleaning - Has the client specific procedures in mind? Are they adequate? - Have similar data been analyzed by the client or others? References? --- ## 6. Recommendations - Data checking and assessing data quality - Data exploration (histograms, box plots, scatter plots, simple linear regres- sions) - Preprocessing: Outlier removal, transformations, data cleaning - Statistical model, hypothesis test, confidence regions - Diagnostics and residual analysis - Computational implementation and software - Possible interpretation of results --- ## 7. Summary/ Conclusions - Additional consulting schedule if applicable - Who will do what? - Computing issues - Timeline --- ## Acknowledgement Slide contents: 6-12 Hans-Georg Müller Professor at the Department of Statistics, University of California, Davis Web: https://anson.ucdavis.edu/~mueller/