What is Trend Analysis in Aviation SMS
Simply put, trend analysis in aviation SMS is an activity that is used to target patterns in your safety program that might otherwise be “lost in the noise” of your safety data.
Patterns can be related to behavior, occurrences, or other aspects of your safety program. Patterns are identified within your chosen time period.
You may identify trends that are:
- Yearly; or
- Over multiple years.
As trends are identified over a period of time, trend analysis is usually linear. However, it is possible to identify trends that are non-linear.
Trend analysis is one of the simplest, but most powerful methods for data analysis. It is an essential reactive risk management and Safety Assurance activity. Here is how to perform trend analysis in aviation SMS.
1 – Choose Which Pattern You Want to Identify
The first and most obvious step in trend analysis is to identify which data trend you want to target. For example, you could look for trends in:
- Number of reported issues in a department, over time;
- Number of lost time injuries in company, over time; or
- Number or high risk, initials risk assessments over time.
There are many trends you can choose to target. It’s simply a matter of reviewing the data that you have collected, and choosing which data to hone in on.
2 – Choose Time Period
Trends are analyzed within a given time frame. To target a trend, you need to start by deciding which time frame best fits the trend you are attempting to identify. When identifying trends, it’s important to consider your “data points” within those trends.
Data points are simply the “markers” within the time period where you will identify data. For example:
- Yearly trend: you might have data points for each month;
- Monthly trend: you might have data points for each day;
- Weekly trends: you might have data points for each day; and
- Seasonal trends: you might have data points for each week.
So, when deciding your time period, also decide how specific you want your data points to be.
3 – Choose Types of Data Needed
Based on pattern and time period you chose, you need to choose the specific data that you need. This includes choosing:
- The specific data related to your targeted pattern; and
- That specific data within the given time period.
In other words, you only want to gather the essential information. If you are identifying the number of reported issues in a department for March, 2017, the three data points you need are:
- Reported Issues; and
- The month of March.
This seems like a self-evident step, but in more complex trend analysis this step is very important for verifying the accuracy of your data.
4 – Gather Data
After choosing the types of data needed, simply gather your data. In automated aviation safety databases, this is very simple because the data is gathered automatically. You simply need to apply filters to your data.
In manual safety programs using point solutions or Excel, this data will need to be gathered manually. Depending on the level of organization and detail, this can be easy or hard.
The purpose of gather data is to filter out all information that is not relevant to your trending target.
5 – Use Charting Tools to Visualize Data
With your data hand, you should chart this data with trending charts. Usually this happens by charting this data to a linear line graph:
- X axis is time, with markers being you’re your time period’s data points; and
- Y axis is the number occurrences for the data pattern you are identifying.
In automated SMS, these charts are created for you. In manual systems, you will have to plot these charts manually.
6 – Identify Trends
Identifying trends is relatively simple. Either you see a pattern in the data, or you don’t.
The pattern could be a noticeable:
- Increase over time;
- Decrease over time; or
- Predictable peaks/valleys at certain points on the graph.
Final Thought: Advanced Trend Analysis
Trend analysis becomes more complex when you perform trend analysis by comparing one trend against another in order. This can result in the identification of a more complex trend, such as the relationship between lost time injury trend data and employee turnover trend data.
Advanced trend analysis data is akin to aviation leading indicators. However, you should be careful in performing advanced trend analysis because, while it is very powerful, it can be prone to error and false conclusions – correlation does not equal causation.
Here are several resources that are very helpful for trend analysis: