The Use of Technology in Data Collection and Analysis

Any three of:

• Geospatial systems (GIS/GPS)
• Satellite sensors (remote sensing)
• Radio tracking
• Computer modelling
• Crowd sourcing (citizen science)

Big data means data sets that are so large, varied and fast-generated that ordinary tools cannot store or analyse them.

It needs special computer tools and processing power to handle, and is often described by its huge volume, variety and velocity (speed of generation).

Satellite sensors (remote sensing): sensors carried on satellites detect and record data from a distance without touching the surface. They are used to monitor large or remote areas — for example tracking ice-sheet melt, sea-surface temperature, or forest loss across whole continents — and can do so repeatedly over time to show trends.

Radio tracking: a small radio tag is fitted to an animal, which sends out a signal that can be picked up by a receiver. This lets scientists follow the animal's movements — for example mapping the migration route or territory of a bird, turtle or large mammal — which could not be done by watching it directly.

Benefits:

• Wide coverage and pattern-spotting: the huge amount and variety of data lets vast or remote areas be monitored and reveals patterns and trends that small samples would miss.
• Real-time monitoring: because new data are generated very fast, changes such as a forest fire, pollution event or flood can be detected and responded to quickly.

Limitations:

• Cost and capacity: storing and processing such large, fast-arriving data sets is expensive and needs powerful computers, storage and specialist expertise.
• Trustworthiness: data quality varies (faulty sensors, cloud cover, untrained crowd-sourcing), so errors can enter and spread through the analysis — more data is not the same as better data.

Amount and type of data stored. Benefit: very large, varied data sets allow vast or remote areas to be covered and many types of data (images, sensor readings, locations) to be combined for a fuller picture. Limitation: storing and processing so much varied data is expensive and needs large capacity, and a bigger data set is not automatically more useful — it can contain irrelevant 'noise'.

Speed at which new data is generated. Benefit: the high velocity allows real-time monitoring, so events such as fires or pollution can be detected quickly, and continuous records build up to show long-term trends. Limitation: data may arrive faster than they can be checked or analysed, so important signals may be missed and backlogs build up.

Trustworthiness of the data. Benefit: a very large sample reduces the effect of single anomalies, and automated sensors can avoid some human error. Limitation: data quality varies (faulty or mis-calibrated sensors, cloud cover, untrained volunteers in crowd-sourced records), so it can be hard to verify, and errors can spread through the analysis.

Ways the data can be used. Benefit: powerful analysis can spot patterns and trends, feed predictive computer models, and be shared and mapped to guide management and conservation. Limitation: analysis needs specialist expertise and software not all organisations can afford, there are privacy and ownership concerns, and the data can be misused or misinterpreted (correlation is not causation).

Introduction. Technologies such as geospatial systems, satellite sensors, radio tracking, computer modelling and crowd sourcing have transformed how environmental data are collected, generating enormous 'big data' sets. But collecting more data, over wider areas and faster, is not the same as collecting more reliable data. Whether technology has made data collection 'far more reliable' depends on a clear criterion: the quality of the technology and how carefully the data are checked, rather than simply how much data are gathered. This essay weighs the ways technology has improved reliability against the reasons it does not guarantee it.

The case that technology has made data more reliable.

• Satellite sensors and GIS cover vast and remote areas consistently and repeatedly, so monitoring no longer depends on a few hand measurements that might be unrepresentative.
• Automated sensors record continuously and can be calibrated, avoiding some kinds of human error and fatigue.
• Large samples mean a single anomalous reading has less effect, so genuine patterns and long-term trends (such as warming or forest loss) stand out more clearly.
• Data can be stored, re-checked and shared, so results can be verified by others — a key part of reliability.

The case that technology has not automatically made data reliable.

• Technology can still fail or be biased: sensors can be faulty or mis-calibrated, satellites are blocked by cloud cover, and computer models are only as good as their assumptions.
• Crowd-sourced data depends on untrained volunteers who may misidentify species, and records cluster in accessible places, so coverage is uneven and biased.
• 'More data' is not 'better data' — a huge data set can be incomplete or full of noise, and errors in the inputs can spread through the analysis.
• Big data can be misinterpreted or misused (a correlation in a large data set does not prove causation), and proper analysis needs expertise not everyone has.

Judgement. The decisive criterion is whether the data are well collected and verified, not merely plentiful. Where technology is well designed, calibrated and its outputs are checked — for example satellite monitoring of deforestation cross-checked with ground surveys — it genuinely makes data collection far more reliable than old hand methods, because of its coverage, consistency and the ability to repeat and verify. But technology does not make data reliable on its own: faulty sensors, biased crowd sourcing and unverified big data can all mislead. The statement is therefore an overstatement — technology has made data collection far more powerful and wide-reaching, and more reliable when used carefully and checked, but reliability still depends on good method and verification, not on the technology alone. The responsible approach is to use technology's reach while continuing to check data quality and treat very large data sets with appropriate caution.