In my personal experience, what I thought was anxiety largely went away when I was treated for depression.

So I'm just gonna recommend what Scott has to say on that matter:

http://slatestarcodex.com/2014/06/16/things-that-sometimes-help-if-youre-depressed/

Do you do any sports? Martial arts classes for example gives you an environment where you face your anxiety head on.

I can offer at least two point of view.
The first is that what I thought was massive social anxiety was actually just social inexperience, that is a large part of my anxiety derived from not knowing what was the accepted social protocol in a given situation. Usually sitting quietly and observing what others did helped.
The second is that you need to subdivide and identify which steps of social interactions you are able to do and which you aren't. For example, instead of just throwing yourself into a social gathering, you can (for example) get ready and go out from your house, but not get in front of the place. Or you can get in front of the place but not enter. Or you can enter but you have a sense of urgency that prompts you to leave immediately after, etc. Instead of "just practice" the whole interactions, identify the smallest next step that you can practice, and if you can't practice that step, subdivide into even smaller units (e.g. literally just doing the next step).

I recommend reading section 19 (on the management of social anxiety disorder) in the recent treatment guidelines from the British Association for Psychopharmacology (pp. 17–19). A sample:

19.1. Recognition and diagnosis

Social anxiety disorder is often not recognised in primary medical care (Weiller et al., 1996) but detection can be enhanced through the use of screening questionnaires in psychologically distressed primary care patients (Donker et al., 2010; Terluin et al., 2009). Social anxiety disorder is often misconstrued as mere ‘shyness’ but can be distinguished from shyness by the higher levels of personal distress, more severe symptoms and greater impairment (Burstein et al., 2011; Heiser et al., 2009). The generalised sub-type (where anxiety is associated with many situations) is associated with greater disability and higher comorbidity, but patients with the non-generalised subtype (where anxiety is focused on a limited number of situations) can be substantially impaired (Aderka et al., 2012; Wong et al., 2012). Social anxiety disorder is hard to distinguish from avoidant personality disorder, which may represent a more severe form of the same condition (Reich, 2009). Patients with social anxiety disorder often present with symptoms arising from comorbid conditions (especially depression), rather than with anxiety symptoms and avoidance of social and performance situations (Stein et al., 1999). There are strong, and possibly two-way, associations between social anxiety disorder and dependence on alcohol and cannabis (Buckner et al., 2008; Robinson et al., 2011).

19.2. Acute treatment

The findings of meta-analyses and randomised placebocontrolled treatment studies indicate that a range of approaches are efficacious in acute treatment (Blanco et al., 2013). CBT [cognitive behavioral therapy] is efficacious in adults (Hofmann and Smits, 2008) and children (James et al., 2005): cognitive therapy appears superior to exposure therapy (Ougrin, 2011), but the evidence for the efficacy of social skills training is less strong (Ponniah and Hollon, 2008). Antidepressant drugs with proven efficacy include most SSRIs (escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline), the SNRI venlafaxine, the MAOI phenelzine, and the RIMA moclobemide.

[...]

19.4. Comparative efficacy of pharmacological, psychological and combination treatments

Pharmacological and psychological treatments, when delivered singly, have broadly similar efficacy in acute treatment (Canton et al., 2012). However, acute treatment with cognitive therapy (group or individual) is associated with a reduced risk of symptomatic relapse at follow-up (Canton et al., 2012). It is unlikely that the combination of pharmacological with psychological treatments is associated with greater overall efficacy than with either treatment, when given alone, as only one in four studies of the relative efficacy of combination treatment found evidence for superior efficacy (Blanco et al., 2010). The findings of small randomised placebo-controlled studies suggest that the efficacy of psychological treatment may be enhanced through prior administration of d-cycloserine (Guastella et al., 2008; Hofmann et al., 2006) or cannabidiol (Bergamaschi et al., 2011).

From a patient perspective, the guidelines suggest that each of the following four approaches should be similarly effective for the treatment of social anxiety as long as the care provider is adequately trained and up-to-date with current best practice:

• Pharmacotherapy
  • given by a psychiatrist.
  • given by a primary care physician.
• Psychotherapy
  • with a therapist.
  • in a group setting.