Python Tutorial: Importing & exporting data

Want to learn more? Take the full course at https://learn.datacamp.com/courses/pandas-foundations at your own pace. More than a video, you'll learn hands-on coding & quickly apply skills to your daily work. --- Now, let's extend our skills for reading DataFrames from files. We'll use a comma-separated-values file of sunspot observations collected from SILSO (Sunspot Index & Long-term Solar Observations). The entries date back to the 19th century with over seventy thousand rows. The read_csv function requires a string describing a filepath as input. We read into a DataFrame sunspots. Using info, we see the DataFrame has mostly integer or floating-point entries. Notice the index of the DataFrame (the row labels) are of type RangeIndex (just integers). Let's use the accessor dot iloc to view a slice of the middle of the DataFrame. We can see some of the problems: the column headers don't make sense and there are many perplexing negative one entries in one column. What's going on? First, the CSV file does not provide column labels in a header row. The column meanings can be gleaned from SILSO's website. Columns zero through two give the Gregorian date, column three is a decimal value of the date, column four is the number of sunspots observed that day, and column five indicates confidence in the measurement (zero or one). Second, the negative ones in column four denote missing values; we need to take care of those. Finally, as written, the dates are awkward for computation, a common problem with CSV files. Let's tidy this up. Using header equals None prevents pandas from assuming the first line of the file gives column labels. Alternatively, an integer header argument gives the row number (indexed from 0) where column labels actually are and the data begins. Notice, now, the columns & rows are assigned integers from 0 as labels. We can explicitly label the columns with the option names. We define a list of strings col_names to label the columns pr