#coursera

Q: I'm interested in Artificial Intelligence but don't know where to start. Is a Machine Learning course a good option for beginners, and what can I expect to learn?

A: If you're looking for a beginner-friendly way to enter the world of AI, Machine Learning is an excellent place to start. The course explains machine learning concepts in a clear, structured way, making it suitable even if you have little or no prior experience. You'll explore how machine learning models work, understand the basics of data-driven decision-making, and discover how AI is used across industries such as healthcare, finance, retail, and cloud computing.

One of the biggest advantages of Coursera Machine Learning is its practical learning approach. Instead of focusing only on theory, you'll gain hands-on experience with real-world examples that help reinforce key concepts. Whether you're a student, software developer, data enthusiast, or working professional looking to upskill, the course provides a solid foundation for exploring more advanced AI and machine learning topics.

1. Program

import pandas import numpy

bug fix for display formats to avoid runtime errors

pandas.set_option('display.float_format', lambda x: '%f' % x)

load the GapMinder dataset

data = pandas.read_csv('GapMinder_Dataset.csv', low_memory=False)

number of observations (rows) and variables (columns)

print(len(data)) # number of observations (countries) print(len(data.columns)) # number of variables

convert variables of interest to numeric

data['incomeperperson'] = pandas.to_numeric(data['incomeperperson'], errors='coerce') data['alcconsumption'] = pandas.to_numeric(data['alcconsumption'], errors='coerce') data['polityscore'] = pandas.to_numeric(data['polityscore'], errors='coerce')

print("counts for incomeperperson - GDP per capita (PPP, constant 2011 USD)") c1 = data['incomeperperson'].value_counts(sort=False, dropna=False) print(c1)

print("percentages for incomeperperson - GDP per capita (PPP, constant 2011 USD)") p1 = data['incomeperperson'].value_counts(sort=False, normalize=True, dropna=False) print(p1)

print("counts for alcconsumption - litres of pure alcohol per person per year") c2 = data['alcconsumption'].value_counts(sort=False, dropna=False) print(c2)

print("percentages for alcconsumption - litres of pure alcohol per person per year") p2 = data['alcconsumption'].value_counts(sort=False, normalize=True, dropna=False) print(p2)

print("counts for polityscore - democracy/autocracy score (-10 to 10)") c3 = data['polityscore'].value_counts(sort=False, dropna=False) print(c3)

print("percentages for polityscore - democracy/autocracy score (-10 to 10)") p3 = data['polityscore'].value_counts(sort=False, normalize=True, dropna=False) print(p3)

(The full income and alcohol tables run to 191 and 181 rows respectively, since almost every country has a unique value — the excerpts above and full run in the attached script show the pattern; the tables are truncated here for readability but the complete output is reproducible by running the script.)

3. Description of frequency distributions

The GapMinder dataset contains 213 countries and 16 variables. For this stage of the analysis I examined the frequency distributions of my two research variables — income per person and alcohol consumption — plus a third variable, polity score (a measure of how democratic or autocratic a country's government is, ranging from -10 to +10), which I included because it produces a cleanly interpretable table.

Income per person ranges from about $104 to just over $105,000 across countries, and because it's a continuous variable, almost every one of the 190 countries with a value has its own unique income figure — each appearing once( each has 0.47% of the sample as standard). There are missing values which make up 10.8% of the sample, meaning income data was not available for 23 of the 213 countries in the dataset.

Alcohol consumption ranges from 0.03 to 23.01 litres of pure alcohol per person per year, and shows the same pattern — almost entirely unique values, each occurring about 0.47% for each observation of the sample . There are 26 missing values which is 12.2% of the sample set.