Question • Is there…. What is… • Make sure your research helps the study, help you understand the data e.g. http://archive.ics.uci.edu/ml/datasets/Abalone • Make sure you add variables, units and the population • use 1 variable then maybe split it e.g. male and female, to look at data more clearly and extend study.
Display Information • Labels with units • Just a word doc with images attached
Features • Overall (direction, strength) • Outliers – where are they? What would happen if you took them out? • Groupings • Where in the graph does data increase rapidly etc.
Relationship • Visual – talk about what you see without the model fitted!! • Strength – how close to trend points (use visual inspection) are then back up with correlation coefficient • Nature of relation in context – full sentence with variables! • Correlation/causation – discussion… • In this case – excellent to talk about the different species and fact the data was collected for science i.e. Paua were killed and collected legally for science so smaller paua caught – relevant to fisherman… • http://www.fish.govt.nz/en-nz/Recreational/Most+Popular+Species/Paua/default.htm • Some of this could be included in conclusions • Breading age could also be handy
Appropriate Model • Fit a trend line to the data • Linear or not linear? Do you need to split the data? • Does it fit? • Will it give you an accurate prediction? • Is it more fitting for some parts of the data and not others? • Could the model be improved?
Prediction • Give two predictions for the response variable – interpolation and extrapolation • Substitute for the x value, use appropriate rounding. • How precise is it? Based on statistical evidence? E.g. fit of trend line • Justify choice of variable • Reflect and what population would it be relevant for? • The prediction may have been made using alternative models and the accuracy of these has been discussed in the context of the investigation.
Conclusion • For example, clearly communicated each component of the cycle. They have extended their initial investigation, use of different charts and graphs, giving contextual reasons to justify the extension and discussing the results of the extended investigation with respect to their question. • Reflection and improvement