There are there levels of trading automation existing nowadays:
1. Screening. This is something like advanced filters, when the system conduct search for specified setup and inform trader when in takes place, so the trader need to make a decision.
2. Semi-automated trading - the system not only find setup, but also open the position according to trading rules. In such case, the system has to calculate the risk per trade and proper position size, while the trader would watch position and close it manually. Similiar principle is the most wiely spread, as it is used by institutional players when they need to get relatively large position. They just set the price and computer would buy or sell as close to that price, as possible.
3. Automated trading, when system itself open and close trades without any human interference.
Most of such tools are based on Python, which is the best solution for this purpose. In fact, you can use Python scripts to do almost anything including analysis and backtesting, but it coul be difficult for most of the traders so they would rather use special software with graphical UI (for example, Forex Tester for backtesting).
If to speak about the fundamental analysis, the computers are widely used to collect and compare data from the earnings reports. For exapmle, you can download data for 2 years (4x2 quaterly reports) and compare main metrics. Computers also could perform contextual analysis calculating the number of positive and negative words to define whether the news are positive or negative. These are the most complicated systems, and they could possibly improve with the development of AI and neural networks.
