DataScience

Sound Story and Presentation - Can you perform Hypothesis Testing - Your blog (better Github and blog) should contain code, visualisations and sufficient text Statistical Analysis - Describe your data and Hypothesis - Buiding a Model with (regression, decision tree vs.) - Verify your Model (it does not have to be successful) Machine Learning Techniques Use at least two of them You can check our blog and learn about the project and what we have done so far. [login to view URL] The links to the dataset are present there. Below there is a comment from the assistant concerning what we have done so far in our blog and his further requests: - The requirement of more than one data source may not be because of credibility concerns only, please check with your instructor. I think you are being forced to be more creative to combine and correlate different data from multiple sources. It is probably not enough to say "energy consumption in both sectors has increased". Your audience may respond: "So what?". Try to tell an interesting story using statistics, so maybe you should find some complementary data to relate to, like economic indicators. - It is good that you now have plotted some graphs but they are not quite self-explanatory. For example, what is the unit of the vertical axis? The source you provided mentions "Btu" but you do not say what it actually means, like 1 Btu = 1055 Joules. Are your graphs of "Total Energy" or "Primary Energy"? State it clearly and explain. Also define what commercial sector and industrial sector is. I do have an idea but the two concept somewhat overlap in my mind so you should not make anybody guess. - The scatter plot looks interesting. The time axis have been eliminated but could be shown as labels of the circles if necessary. Visually it looks as if there would be a positive correlation coefficient between "Commercial" and "Industrial" but now there may be multiple points for some x-coordinate values, violating the functional form y=f(x). You can not fit a function like this if you wanted to. Maybe you could go after x = g(y).