As you can see even though there were more people with diarrhoea in town B, the prevalence of diarrhoea was less in Town B than in Town A. 
While prevalence provides you an idea of the probability or the likelihood of a disease, incidence provides you with the idea as to how fast the disease is developing, or an idea as to the hazard associated with the disease. The word hazard here indicates the chance that someone who is disease free up till now will develop the disease or the health condition under study. For this, we need to start with individuals who are free from the disease under study and a sense of time through which they are followed. This is the concept of person-time. If I person is followed for 1 year, we have 1 person-year. If 10 persons are followed for 1 year, we have 10 person-years. If 100 persons are followed for 1 year, we have 100 person-years. This is same as if 10 persons were to be followed for 10 years, as there would be 10 persons * 10  years = 100 person-years. This will be the denominator and on the numerator we will estimate the number of new cases of the disease. Such as measure is referred to as incidence density rate or simply incidence rate per 1000 person-years or how many person-years we consider. 
Imagine you want to study a study heart diseases among young people (5-16 years of age) in a population P for five years. Let us say it is a small community of 1000 young people, so P = 1000. In the five  years you study them, 50 children were diagnosed with heart disease. No one died in the community and so you have 5000 person-years (1000 people followed for five years will be 5000 person-years). Can you estimate the incidence rate of heart disease in this community?