Скачать или смотреть Data + Geospatial Data Trainings Recap session

Introduction to Data Set Exploration Niyobyose Branchal introduced basic data exploration functions in Python for working with a data set, including how to view the first five or a specified number of observations using `head()` and the last observations using `tail()`. They also demonstrated how to check the shape of the data set, which revealed 5,000 observations/rows and 13 columns, and how to print the list of column names. Branchal emphasized the importance of studying variables, including their data type and shape, as part of data analysis.
Descriptive Statistics of Numerical Variables Branchal showed how to use the `describe()` function to get a summary of numerical columns in the data set, providing insights like count, mean, standard deviation, minimum, and maximum values. For example, they found the youngest person was 18 and the oldest was 24, and the mean total score was 75 with a maximum of 99.99. Branchal noted that this function is very important for summarizing numeric variables and getting initial insights.
Subsetting Data and Column Selection Branchal explained how to select or "subset" specific columns from the data set using single or double square brackets for multiple columns, noting that Python is case-sensitive, which requires the exact column names. They demonstrated selecting columns like 'gender', 'age', 'department', and 'quizzes average'. Branchal then clarified that selecting is choosing columns, while filtering requires meeting a specific condition on the rows.
Filtering Data and Logical Operators Branchal demonstrated filtering data by comparing values, which produces a series of logical (True/False) values, and then subsetting the data to only include rows where the condition is True. They showed how to use the `unique()` function to identify distinct values in a column, confirming the gender categories were 'female' and 'male'. Branchal also illustrated filtering for numerical conditions, such as ages greater than or equal to the maximum age, and combining multiple filter conditions using the logical operators "and" (`&`) and "or" (`|`). Branchal confirmed that the double equal sign (`==`) is used for comparison, while a single equal sign (`=`) is for assignment.

Data Visualization with Pie Charts and Bar Charts Branchal shifted to data visualization, first demonstrating how to create a pie chart to visualize the proportion of gender in the data set using the data's value counts and the `plot(kind='pie')` function. They explained how to add percentages to the pie chart using the `autopct` argument and how to use the Matplotlib library (imported as `plt`) to set the title and remove the count label. Finally, Branchal showed how to use the Seaborn library (imported as `sns`) to create a count plot, or bar chart, to visualize the counts of departments across gender, setting a title and Y-axis label for frequency.
Comparison of Multiplotlib and Seaborn Niyobyose Branchal and Alexis Niyitegeka discussed the differences between Multiplotlib (Mpro) and Seaborn (Sibon), noting they perform almost the same job. Niyobyose Branchal explained that Mpro is more complicated than Seaborn, and while Seaborn uses "axis" to plot, Mplotli uses variables. Alexis Niyitegeka also inquired about changing the default colors like blue and red in the plots, to which Niyobyose Branchal confirmed that colors can be changed, although some functions display their own default colors which cannot be changed via arguments like `colors`..
Using the Explore Function for Dynamic Mapping Niyobyose Branchal introduced the `explore` function from `geopandas` for creating dynamic maps, which allows for zooming and displaying values as one scrolls. They attempted to demonstrate plotting a map of Rwanda's districts, specifying a column like `shape_area` and setting `legend=True` to show color variations based on the data values. Niyobyose Branchal noted that their environment (Google Colab and VS Code) was not working properly to render the dynamic map, and instead showed an example of how the `data.explore()` function, using a `population` column, would display a legend where large populations appear red and smaller populations appear blue.
Questions and Session Wrap-up Byukusenge Jean Paul asked about how to get output for presentations, and Niyobyose Branchal replied that one can save figures and tables. Both Byukusenge Jean Paul and Solomon Saye requested that Niyobyose Branchal share the practical example/notebook, and Niyobyose Branchal agreed to share it. Niyobyose Branchal confirmed that Visual Studio Code works offline, and that the suggestions for functions are provided by artificial intelligence, which also works offline. BINAYISA Gad asked about descriptive scripts that they can learn from, and Niyobyose Branchal assured them that they would share the notebook being used.