Choose four of the following six mini projects:
1. A colleague asks you about the comparative advantages and disadvantages of R and
Python for her project on analysing online sales for your website selling books. Please,
write her a 500 words email and recommend either language.
2. Let’s try real-life job interview questions on R. These questions are taken from one of
the many online examples.
- Describe the data structures in R and their differences.
Please note that this is not supposed to be original research but just a summary of
commonalities and differences. Your answer should not be more than 500 words in
total.
- If you have df <- data.frame(a = c(1, 2, 3), b = c(4, 5, 6), c(7, 8, 9)). How do you select
the c(4, 5, 6)? How do you select the 1? What is df[, 3]? How do you select the first
two columns.
3. Choose either a Twitter identifier (@) or hashtag (#) or a public Facebook page to
analyse. Run a query against Twitter to extract 250 Tweets or extract all posts on the
Facebook page. Produce two analyses. The first analysis should showcase the most
influential users and the second should visualise the textual content with a
wordcloud. In particular:
- In Twitter the most influential users are those with the largest number of followers.
So, you need to either find those from a group of user using a # or find the most
influential one for a @.
- For Facebook, the most influential users are those whose posts had the largest
number of likes, comments and shares.
Please, include in your answer at least one API key to demonstrate that you have
created your own and describe in 2-3 sentences your result.
4. Download the Harry Potter books from
https://archive.org/details/Book5TheOrderOfThePhoenix and present a text analysis,
which includes at least one visualisation of frequent terms and also a model of 5
common topics in the collection. Show the terms per topic, please.
Please note that the link above only points to book 5 of the series, but you can find
links to all the Potter books on that page. Your analysis should be based on the TM
package. Do not use the TidyText package, as this is part of the provided report
example. Please describe the visualisation and topics in 3-4 sentences.
5. Use the ggplot2 package to produce 2 types of visuals by plotting the diamonds
dataset
(https://www.rdocumentation.org/packages/ggplot2/versions/2.2.1/topics/diamond
s). One visual should display the diamonds features of your choice plotted as a
scatter plot (geom_point), while the second visual should display other features
plotted on a violin plot (geom_violin).
Next to the code and the visuals include a short (3-4 sentences) explanation of what
you are observing in your answer.
6. You are working in a marketing department for a big online company and are asked
to design an analysis in order to predict revenue increase from future ad campaigns.
You have one sample dataset on a single campaign and quarterly total sales
generated by Newspaper, TV and Online ad campaigns and associated expenditures:
Sales, Newspaper, TV, Online
16850, 1000, 500, 1500
12010, 500, 500, 500
14740, 2000, 500, 500
13890, 1000, 1000, 1000
12950, 1000, 500, 500
15640, 500, 1000, 1000
14960, 1000, 1000, 1000
13630, 500, 1500, 500
Thus, Quarter 1 indicates that £1000, £500 and £1500 were spent on Newspaper, TV
and Online ad campaigns respectively and total sales during that quarter was £16,850.
Your task is to prepare a short executive summary (200-300 words) to motivate
campaign teams in your company to report on their data in a similar way. This
summary should mainly describe the steps and advantages of your predictive analysis
as well as the kind the predictive analysis you plan to do in order to find out the
relationship with sales and advertising expenditure for the campaigns.
Hint: You need to define which feature you would like to predict and which features
are the predictors. Then, think of their data types to narrow down the prediction
models you can apply. Finally, if you want to do the data analysis in R, you can copy
and paste the data above into a spreadsheet, save it as a csv file and load it into R. Or
you can even directly edit data