Tag: pandas
All the articles with the tag "pandas".
-
Data Analysis Study #7 — Busan Port Open Data: I Thought the Data Was Clean Because the Filter Returned All False
I pulled two real public datasets from Busan Port Authority. I scrapped the question I'd set up for the container traffic statistics midway through, and worked with encoding and berthing time for the vessel arrival/departure data. The outlier filter came back all False because the comparison wasn't valid to begin with.
-
Data Analysis Study #6 — I gave axis=0 in pandas but got column totals
I went over pandas' axis concept, then loc/iloc indexing and boolean mask filtering. I memorized axis numbers as row/column, but the result came out opposite and confused me for a while.
-
Data Analysis Study #5 — pandas read_csv took the first row as column names
I started actual analysis with the Iris dataset. As soon as I read the csv, the column names were broken, and I had to fix it with skiprows and header. I cover what to look at in descriptive statistics, and why you copy the original before attaching derived variables.
-
Data Analysis Study #4 — When I Get a pandas DataFrame, I Start with info()
Following Series, I looked at DataFrame. I cover making a table into a dictionary and putting it into a DataFrame, the difference between pulling a column out as a Series versus as a DataFrame, and grasping the shape of a table first with dtypes and info().