Before we delve into the material, there are two important precursors to the logic that shall be put into explaining just why tangent graphs are typically uphill, while cotangent graphs are typically downhill.
Firstly, it's crucial to know that the ratios for tangent and cotangent, in terms of sine and cosine, are sin/cos, and cos/sin, respectively. From that, we would know that whenever cosine is zero, tangent shall be undefined, resulting in an asymptote, and the same goes for cotangent whenever sine is zero as well.
Secondly, we should know that the x-axis for these graphs is in terms of the unit circle - of π and all that. And, we know that in certain areas of the unit circle, tangent/cotangent are positive/negative. They're both positive in Quadrants I and III, and therefore, are negative in Quadrants II and IV. Since every π/2 on the x-axis is a quadrant, the areas where tangent/cotangent are positive/negative can be organized neatly into certain portions of the x-axis.
Basically, asymptotes divide these portions into repeating portions of either positive to negative (going from left to right), or negative to positive. The reason for which tangent and cotangent differ in where they are allowed to be positive/negative, is that their asymptotes are located in different places - where sine or cosine are zero.
In the above picture, there are asymptotes at π/2 and 3π/2, since that is where cosine is equal to zero. In the sections where tangent is allowed to continually exist for two sections of the graph, it goes from negative to positive, from left to right.
In the above picture, there are asymptotes at π and 2π, since that is where sine is equal to zero. In the sections where cotangent is allowed to continually exist for two sections of the graph, it goes from positive to negative, from left to right.