In this sense, econometrics transforms the discipline of theoretical economics into policy and decision-making tools in the public and private sectors.
Evaluation of the causal effects in an econometric model. This involves specification testing, estimation, and evaluation of results. Most of the time in econometrics your data is observed and not the result of a properly designed random sample or experiment. In such cases, any causal results are subject to the correctness of your model and the econometric analysis does not prove anything. Data mining or the examination of a variety of models and choosing the best fitting will lead to spurious results
Forecasting. You do not need to start with an econometric model although you can use economic theory to formulate your forecasting model. No matter how good your forecasting model is, you can not infer causal relationships from your data
Data reduction or description. You can do anything sensible here but again causal effects can not be inferred.
What are you trying to achieve? Do you have an economic model that you are evaluating? These are important questions and you must have answers before you do your econometric/statistical analysis.
You have 12 years of data. If the data is annual you will probably be restricted to data reduction or description. If your data are quarterly or monthly you still may not have enough data. This depends on your data. Are your data stationary or non-stationary?
Econometrics is not about getting a data set and experimenting with various menus in your econometric software. This is a sure way to spurious results.
Can I recommend Huntington-Klein (2022). The Effect, CRC, which is available online at https://theeffectbook.net/. This explains what should happen before you apply your econometric recipes
First, see whether the data is univariate or multivariate. If you are using a regression model you have to check whether the data is stationary or not. Stationarity tests are important. If the data set is not stationary one can transform it by using a difference equation, logarithm, etc. A time series generated is thought of as a stochastic process. The test can be based on a correlogram, a unit root test of stationarity, etc. The underlining assumption of time series data is that the data is stationary. A stochastic process is stationary if the mean and variance are constant over time.
Econometrics is statistics applied to economic problems. It can never replace your thinking or theory, respectively. That means, you have to specify your model/equation (=hypothesis) before you try econometric estimations (=test). The choice of the econometric method is mainly dependent on your specification, but of course on the disposible data, too.