Connecting the Dots: Making Sense of Data

Five friends go fruit-picking and collect 8, 10, 12, 14 and 6 mangoes respectively. If they tipped all the mangoes into one basket and shared them out equally, how many would each person get? Adding the amounts gives 8 + 10 + 12 + 14 + 6 = 50 mangoes in total, and sharing 50 mangoes among 5 friends gives 50 ÷ 5 = 10 mangoes each. This equal-share number, 10, is the arithmetic mean (or simply the average) of the five amounts, even though not one single friend actually picked exactly 10 mangoes.

In general, Mean = (Sum of all the values) ÷ (Number of values). The mean is useful because it reduces a whole group of numbers to one representative figure, which makes comparing two different groups much easier — comparing two totals directly can be misleading if the groups are different sizes, but comparing two means puts them on equal footing.

This fair-share idea connects the mean to something very physical: imagine levelling out several unevenly filled glasses of water by pouring between them until every glass has the same amount — the final shared level is the mean of the original amounts.

The median is found by sorting the data and picking the middle value. Eight students score 12, 15, 18, 20, 22, 25, 28 and 30 marks in a quiz. Sorted, they are already in order: 12, 15, 18, 20, 22, 25, 28, 30. Since there are 8 values (an even number), there isn't a single middle value, so the median is the average of the two middle values, 20 and 22: (20 + 22) ÷ 2 = 21. If there had been an odd number of values instead, the median would simply be the one value sitting exactly in the middle once sorted.

The mode is simply the value that occurs most often. In a survey of 10 students' weekly pocket money (in rupees): 50, 60, 50, 70, 50, 80, 60, 50, 90, 60 — sorting gives 50, 50, 50, 50, 60, 60, 60, 70, 80, 90. The value 50 appears four times, more than any other value, so the mode is 50. Compare this to the mean of the same data: the total is 620, and 620 ÷ 10 = 62 — noticeably higher than the mode of 50, because a few larger amounts (80 and 90) pull the mean upward even though most students actually received 50 or 60.

A dataset can have more than one mode (if two or more values tie for the highest frequency), or no mode at all (if every value is different).

Consider five heights (in cm): 148, 149, 150, 151, 152. The mean is (148+149+150+151+152) ÷ 5 = 750 ÷ 5 = 150, and sorting gives a median of 150 as well — mean and median agree closely because every value sits close together.

Now suppose one of those five heights was actually a typing mistake, and the real dataset is 148, 149, 150, 151 and 110 (instead of 152) — an unusually low value compared to the rest. The mean becomes (148+149+150+151+110) ÷ 5 = 708 ÷ 5 = 141.6, noticeably pulled downward. But the median, found by sorting (110, 148, 149, 150, 151) and picking the middle value, is still 149 — barely affected at all. The value 110 here is called an outlier: a value that differs sharply from the rest of the data.

This is exactly why outliers matter: the mean takes every single value into account in its sum, so one extreme value can shift it noticeably, while the median only cares about which value sits in the middle position, so it stays far more stable unless the outlier itself happens to be near the middle. When a dataset has a clear outlier, the median is often the more trustworthy “typical value” to quote.

A bar graph uses a bar's height to show the value for each separate category, with categories (like months, or favourite subjects) marked along the horizontal axis and the scale marked along the vertical axis. Reading a bar graph always starts the same way: check the scale (what does one unit of height represent?), then compare bar heights to compare categories.

A double bar graph (also called a clustered bar graph) places two bars side by side for every category, usually in two different colours, so two related datasets can be compared directly at a glance — for instance, comparing the number of visitors to two different parks across the months of a year, with one colour per park.

A histogram looks similar to a bar graph but is used for continuous numerical data grouped into equal-width class intervals (for example, how many students have heights between 140–145 cm, 145–150 cm, 150–155 cm, and so on). Because the data is continuous — height doesn't jump straight from 144 cm to 145 cm, it flows through every value in between — the bars of a histogram are drawn touching each other with no gaps, unlike a bar graph where the bars usually have small gaps between separate, unrelated categories.

A pie chart shows how a total is split into parts, using a circle divided into slices — each slice's size (its angle, out of the full 360° circle) is proportional to the percentage share that category represents. Pie charts are best when the point is to show proportions of a whole (like what percentage of a class prefers each sport), rather than exact raw counts, which a bar graph shows more precisely.

Suppose two players' scores across different numbers of matches need to be compared fairly: Player X scores 10, 25, 30 and 15 runs across 4 matches (total 80, mean 20), while Player Y scores 5, 40, 10, 20 and 5 runs across 5 matches (total also 80, but mean 16). Comparing totals alone would wrongly suggest they performed equally well, but since they played different numbers of matches, comparing their means (20 versus 16) is the fairer way to judge consistency of performance — Player X's mean is higher despite an identical total.

Before trusting a mean, it is worth checking the data for outliers, since one unusually high or low value can distort the picture (as seen with the height example earlier). When outliers are present, reporting the median alongside the mean — or instead of it — often gives a more honest summary.

Finally, always ask what story the chosen graph or summary number is actually telling, and what further questions it raises. A single mean, median or mode is a useful starting point, but genuinely understanding data usually means looking at the spread of values too, not just one central number.