In order to become a successful data scientist, it is important not to jump the gun and directly development into statistical models and Python/R commands which could simply take inputs and provide the model output. Rather, the process should be more organic and synchronous wherein start with preparing the basics of mathematics & statistics and then progress towards finer details.
In addition to this, the approach to solving problems is another important aspect. Simply solving problems would not help until it has a business flavour to it.
Happy Learning :)